Results 1 - 10
of
12
Crowd++: Unsupervised speaker count with smartphones,”
- in ACM UbiComp,
, 2013
"... ABSTRACT Smartphones are excellent mobile sensing platforms, with the microphone in particular being exercised in several audio inference applications. We take smartphone audio inference a step further and demonstrate for the first time that it's possible to accurately estimate the number of p ..."
Abstract
-
Cited by 8 (3 self)
- Add to MetaCart
(Show Context)
ABSTRACT Smartphones are excellent mobile sensing platforms, with the microphone in particular being exercised in several audio inference applications. We take smartphone audio inference a step further and demonstrate for the first time that it's possible to accurately estimate the number of people talking in a certain place -with an average error distance of 1.5 speakers -through unsupervised machine learning analysis on audio segments captured by the smartphones. Inference occurs transparently to the user and no human intervention is needed to derive the classification model. Our results are based on the design, implementation, and evaluation of a system called Crowd++, involving 120 participants in 10 very different environments. We show that no dedicated external hardware or cumbersome supervised learning approaches are needed but only off-the-shelf smartphones used in a transparent manner. We believe our findings have profound implications in many research fields, including social sensing and personal wellbeing assessment.
Lightweight Neighborhood Cardinality Estimation in Dynamic Wireless Networks
"... Abstract—We address the problem of estimating the neighborhood cardinality of nodes in dynamic wireless networks. Different from previous studies, we consider networks with high densities (a hundred neighbors per node) and where all nodes estimate cardinality concurrently. Performing concurrent esti ..."
Abstract
-
Cited by 1 (0 self)
- Add to MetaCart
(Show Context)
Abstract—We address the problem of estimating the neighborhood cardinality of nodes in dynamic wireless networks. Different from previous studies, we consider networks with high densities (a hundred neighbors per node) and where all nodes estimate cardinality concurrently. Performing concurrent estimations on dense mobile networks is hard; we need estimators that are not only accurate, but also fast, asynchronous (due to mobility) and lightweight (due to concurrency and high density). To cope with these requirements, we propose Estreme, a neighborhood cardinality estimator with extremely low overhead that leverages the rendezvous time of low-power medium access control (MAC) protocols. We implemented Estreme on the Contiki OS and show a significant improvement over the state-of-the-art. With Estreme, 100 nodes can concurrently estimate their neighborhood cardinality with an error of ≈10%. State-of-the-art solutions provide a similar accuracy, but on networks consisting of a few tens of nodes and where only a fraction of nodes estimate the cardinality concurrently.
ABSTRACT OF THE DISSERTATION Learning Human Contexts through
, 2014
"... Learning human contexts is critical to the development of many applications, ranging from healthcare, business, to social sciences. Most existing work, however, acquires contextual information in an obtrusive manner – they may require subjects to carry mobile devices, or rely on self or peer report ..."
Abstract
- Add to MetaCart
(Show Context)
Learning human contexts is critical to the development of many applications, ranging from healthcare, business, to social sciences. Most existing work, however, acquires contextual information in an obtrusive manner – they may require subjects to carry mobile devices, or rely on self or peer report to report data. In this dissertation, we present two unobtrusive techniques that can help us learn important human contex-tual information including count, location, trajectory, and speech characteristics. We first present SCPL, a radio frequency-based device-free localization technique. SCPL is able to count how many people are in an indoor setting and track their locations by observing how they disturb the wireless radio links in the environment. Second, we present Crowd++, a smartphone-based speech sensing technique, which records a conversation and automatically counts the number of people in the conversation with-out prior knowledge of their speech characteristics. Both techniques are unobtrusive, low-cost, and private, which can thus enable a large array of important applications that rely upon the knowledge of human contextual information. ii
Crowdsensing the Speaker Count in the Wild: Implications and Applications
"... Abstract-The Mobile Crowd Sensing (MCS) paradigm enables large-scale sensing opportunities at lower deployment costs than dedicated infrastructures by utilizing the large number of today's mobile devices. In the context of MCS, end users with sensing and computing devices can share and extract ..."
Abstract
- Add to MetaCart
(Show Context)
Abstract-The Mobile Crowd Sensing (MCS) paradigm enables large-scale sensing opportunities at lower deployment costs than dedicated infrastructures by utilizing the large number of today's mobile devices. In the context of MCS, end users with sensing and computing devices can share and extract information of common interest. In this article, we examine Crowd++, a MCS application, which accurately estimates the number of people talking in a certain place through unsupervised machine learning analysis on audio segments captured by mobile devices. Such a technique can find application in many domains, such as crowd estimation, social sensing, and personal well-being assessment. In this article, we demonstrate the utility of this technique in the context of conference room usage estimation, social diary, and social engagement in a power efficient manner followed by a discussion on privacy and possible optimizations to Crowd++ software.
Human Object Estimation via Backscattered Radio Frequency Signal
"... Abstract—In this paper, we propose a system called R # to estimate the number of human objects using passive RFID tags but without attaching anything to human objects. The idea is based on our observation that the more human objects are present, the higher the variance in the RSS values of the tag b ..."
Abstract
- Add to MetaCart
(Show Context)
Abstract—In this paper, we propose a system called R # to estimate the number of human objects using passive RFID tags but without attaching anything to human objects. The idea is based on our observation that the more human objects are present, the higher the variance in the RSS values of the tag backscattered RF signal. Thus, based on the received RF signal, the reader can estimate the number of human objects. R# includes an RFID reader and some (say 20) passive tags, which are deployed in the region that we want to monitor the number of human objects, such as the region in front of a painting. The RFID reader periodically emits RF signal to identify all tags and the tags simply respond with their IDs via C1G2 standard protocols. We implemented R # using commercial Impinj H47 passive RFID tags and Impinj reader model R420. We conducted experiments in a simulated picking aisle area of the supermarket environment. The experimental results show that R # can achieve high estimation accuracy (more than 90%). I.
Passive, Privacy-preserving Real-time Counting of Unmodified Smartphones via ZigBee Interference
"... Abstract—The continuing proliferation of smartphones makes them an effective means to monitor the number of people within an area, for example, to gain insights into customer engagement in retail and to enable an intelligent traffic system in a city. However, current approaches to obtain this inform ..."
Abstract
- Add to MetaCart
(Show Context)
Abstract—The continuing proliferation of smartphones makes them an effective means to monitor the number of people within an area, for example, to gain insights into customer engagement in retail and to enable an intelligent traffic system in a city. However, current approaches to obtain this information are either invasive as they require to continuously run a dedicated smartphone app, or they compromise users ’ privacy by sniffing the MAC addresses of their smartphones. As a consequence, lawyers, authorities, and the population are very skeptical toward adopting such innovative systems. We present DEVCNT, the first system that counts in real-time the number of Wi-Fi enabled smartphones in a non-invasive manner while preserving by design the privacy of the smartphone users. This paper details how DEVCNT detects active Wi-Fi scans performed by smartphones on a ZigBee device, and how DEVCNT uses the number of detected scans to estimate the number of Wi-Fi enabled smartphones. Results from controlled and real-world experiments show that DEVCNT: (i) detects more than 99 % of active Wi-Fi scans even under interference from multiple wireless technologies, (ii) achieves up to 91 % accuracy in the estimated smartphone counts, and (iii) provides meaningful estimates in a real test run involving hundreds of Wi-Fi transmitters. I.
Crowd++: Unsupervised Speaker Count with Smartphones
"... Smartphones are excellent mobile sensing platforms, with the microphone in particular being exercised in several audio in-ference applications. We take smartphone audio inference a step further and demonstrate for the first time that it’s possi-ble to accurately estimate the number of people talking ..."
Abstract
- Add to MetaCart
(Show Context)
Smartphones are excellent mobile sensing platforms, with the microphone in particular being exercised in several audio in-ference applications. We take smartphone audio inference a step further and demonstrate for the first time that it’s possi-ble to accurately estimate the number of people talking in a certain place – with an average error distance of 1.5 speak-ers – through unsupervised machine learning analysis on au-dio segments captured by the smartphones. Inference occurs transparently to the user and no human intervention is needed to derive the classification model. Our results are based on the design, implementation, and evaluation of a system called Crowd++, involving 120 participants in 6 very different en-vironments. We show that no dedicated external hardware or cumbersome supervised learning approaches are needed but only off-the-shelf smartphones used in a transparent manner. We believe our findings have profound implications in many research fields, including social sensing and personal wellbe-ing assessment.
REPC: Reliable and Efficient Participatory Computing for Mobile Devices
"... Abstract—Smartphones and mobile devices have greatly penetrated the daily lives of many people. While participa-tory/pervasive sensing has gained wide adoptions by leveraging various onboard sensors on mobile devices, another powerful resource, the computational power on these mobile devices has bee ..."
Abstract
- Add to MetaCart
(Show Context)
Abstract—Smartphones and mobile devices have greatly penetrated the daily lives of many people. While participa-tory/pervasive sensing has gained wide adoptions by leveraging various onboard sensors on mobile devices, another powerful resource, the computational power on these mobile devices has been less frequently harnessed by researchers and practitioners. To fill this gap, we propose in this work the modeling, analysis, and implementation of participatory computing. Specifically, we propose REPC, a generic randomized task assignment frame-work for the participatory computing paradigm, which guar-antees the overall system performance with close to minimal workload at individual participating devices. To achieve these design objectives, we model the intrinsic relationship between the workload of individual devices and the probability they complete their assigned tasks. Based on our modeling results, we analyze the maximal system capacity for any given participatory computing system and derive the minimal workload for individual participating devices to achieve the overall system performance requirement. We have fully implemented our design on the Android platform and demonstrated its performance through a representative participatory computing application. Extensive experiments and simulation results demonstrate that our design is able to achieve more than 90 % task completion ratios with only 10 % system overhead in practice. I.
REPC: Reliable and Efficient Participatory Computing for Mobile Devices
"... Abstract—Smartphones and mobile devices have greatly penetrated the daily lives of many people. While participa-tory/pervasive sensing has gained wide adoptions by leveraging various onboard sensors on mobile devices, another powerful resource, the computational power on these mobile devices has bee ..."
Abstract
- Add to MetaCart
(Show Context)
Abstract—Smartphones and mobile devices have greatly penetrated the daily lives of many people. While participa-tory/pervasive sensing has gained wide adoptions by leveraging various onboard sensors on mobile devices, another powerful resource, the computational power on these mobile devices has been less frequently harnessed by researchers and practitioners. To fill this gap, we propose in this work the modeling, analysis, and implementation of participatory computing. Specifically, we propose REPC, a generic randomized task assignment frame-work for the participatory computing paradigm, which guar-antees the overall system performance with close to minimal workload at individual participating devices. To achieve these design objectives, we model the intrinsic relationship between the workload of individual devices and the probability they complete their assigned tasks. Based on our modeling results, we analyze the maximal system capacity for any given participatory computing system and derive the minimal workload for individual participating devices to achieve the overall system performance requirement. We have fully implemented our design on the Android platform and demonstrated its performance through a representative participatory computing application. Extensive experiments and simulation results demonstrate that our design is able to achieve more than 90 % task completion ratios with only 10 % system overhead in practice. I.
Ambient Rendezvous: Energy-Efficient Neighbor Discovery via Acoustic Sensing
"... Abstract—The continual proliferation of mobile devices has stimulated the development of opportunistic encounter-based networking and has spurred a myriad of proximity-based mobile applications. A primary cornerstone of such applications is to discover neighboring devices effectively and efficiently ..."
Abstract
- Add to MetaCart
(Show Context)
Abstract—The continual proliferation of mobile devices has stimulated the development of opportunistic encounter-based networking and has spurred a myriad of proximity-based mobile applications. A primary cornerstone of such applications is to discover neighboring devices effectively and efficiently. De-spite extensive protocol optimization, current neighbor discovery modalities mainly rely on radio interfaces, whose energy and wake up delay required to initiate, configure and operate these protocols hamper practical applicability. Unlike conventional schemes that actively emit radio tones, we exploit ubiquitous audio events to discover neighbors passively. The rationale is that spatially adjacent neighbors tend to share similar ambient acoustic environments. We propose AIR, an effective and efficient neighbor discovery protocol via low power acoustic sensing to reduce discovery latency. Especially, AIR substantially increases the discovery probability of the first time they turn the radio on. Compared with the state-of-the-art neighbor discovery protocol, AIR significantly decreases the average discovery latency by around 70%, which is promising for supporting vast proximity-based mobile applications. I.
