Activity recognition along with device position recognition can provide contextual cues suitable to infer user interruptibility and device accessibility. Our system fuses data from accelerometer and multiple light sensors to classify activities and device positions. Previously published results achieve robust activity recognition performance with multiple sensors attached to fixed body positions, a model suitable for use cases such as healthcare and fitness. We achieve comparable activity recognition performance using smartphones placed in unknown on-body positions including pocket, holster and hand. Results obtained from a diverse data set show that motion state and device position are classified with macro-averaged f-scores 92.6% and 66.8% respectively, over six activities and seven device positions. We demonstrate the performance of our classifier with an implementation running on the Android platform, that visitors can try out.

Femtocells are low-power cellular base stations that are typically deployed indoors in residential, enterprise or hotspots settings. Femtocell deployments provide excellent user experience through better coverage for voice and very high data throughputs. In this paper, we focus on 3G UMTS femtocells and analyze the impact of femtocells on idle-mode mobility and UE battery life. Detailed dynamic simulations are presented to quantify the impact of femtocells on the number of cell searches, cell reselections and location area updates for both femtocell users and macro users. It is shown that femtocells reduce the number of intra-frequency and inter-frequency searches for femtocell users while increasing the number of searches for macro users. In addition, methods for facilitating capacity offload to femtocells are presented. In particular, cell reselection parameter optimization, use of beacons and enhancements of UE search algorithms are described as possible methods for facilitating capacity offload to femtocells. The tradeoffs between capacity offload and UE battery life are also evaluated.

We present an overview of the MAC and PHY layer enhancements required to ensure high application layer throughput for IEEE 802.11n. We describe an enhanced scheduled mode operation which is shown, through simulation, to yield an application layer throughput that is greater than 100 Mb/s in practical scenarios. We also demonstrate that the proposed enhancements significantly improve the application layer throughput with EDCA.

A method for predicting packet error rates in MIMO-OFDM WLAN systems is presented. The method is based on using post-detection SNRs as an abstraction of the physical layer, and is motivated by the need for a simple and efficient way of modelling the physical layer in system level simulation scenarios involving multiple stations. The physical layer abstraction is sufficient for generating error processes in the system simulations that accurately reflect the interaction between the MIMO-OFDM physical layer and the underlying wireless channel. We validate the abstract model by comparing packet error rates predicted by the model with packet error rates obtained through full link simulations for two different approaches to MIMO processing, referred to as spatial spreading and eigenvector steering.

This paper presents two transmission techniques for MIMO OFDM communication, referred to as spatial spreading and eigenvector steering. Spatial spreading is used when the transmitting station is not presumed to have sufficient channel state information to compute optimum steering vectors. This situation may occur for a variety of reasons, including poor or aged channel estimates or lack of calibration between the transmit and receive antenna chains. Eigenvector steering is used in cases where the transmitting station has sufficient information about the channel to compute optimum transmit steering vectors. Simulation results showing throughput and range achieved by the two transmission strategies are provided.

Radio link control (RLC) protocols are typically employed for reliable in-sequence delivery of service data units (SDUs) in wireless packet data systems. The RLC layer segments packets obtained from the upper layer (referred to as SDUs) into smaller RLC transmission units (or blocks) and uses selective-repeat automatic repeat request (SR-ARQ) for error recovery of RLC blocks. In earlier work, SR-ARQ performance is typically characterized in terms of the long-term throughput or in-sequence delivery delay of RLC blocks. The SDU delivery delay which is a more meaningful measure of RLC performance (in terms of the service provided to a higher layer, e.g., transmission control protocol) has not been quantified. In this paper, we analyze the SDU delivery delay of SR-ARQ as a function of the SDU size and the channel coding scheme employed. Closed-form delay expressions as well as approximations are provided. The analysis is verified through enhanced general packet radio service RLC simulations. Based on the analysis, we propose that link adaptation be backlog dependent in order to reduce the SDU delivery delay at the RLC layer.

In this paper, we propose an efficient scheme to transport video over wireless networks, specifically cdma2000® 1x. Speech transmission over cdma2000® uses a variable rate voice coder (vocoder) over a channel with multiple fixed rates. We apply these ideas to compressed video transmission over wireless IP networks. Explicit Bit Rate (EBR) video compression is designed to match the video encoder output to a set of fixed channel rates. We show that in comparison with VBR video transmission over a fixed rate wireless channel, EBR video transmission provides improved error resilience, reduced latency and improved efficiency.

In wireless packet data systems such as the Enhanced General Packet Radio Service (EGPRS), selective ARQ is used for error recovery over a radio link. In earlier work, the radio link control (RLC) performance is typically characterized in terms of the throughput and the size of the higher layer protocol data units (PDUs) or the packets to be transferred is ignored. We propose that the link adaptation be based on user perceived delay instead of the throughput.

The General Packet Radio Service (GPRS) was introduced by ETSI to meet the ever-increasing demand for wireless data over the GSM radio interface. Enhanced GPRS (EGPRS Phase 1) was later introduced in order to increase the spectrum efficiency and data throughput by using 8-PSK modulation and link adaptation techniques. While (E)GPRS uses packet-switched technology to efficiently transmit best-effort data, real-time services (e.g. voice) are currently supported only via circuit-switched systems. This service separation reduces the potential spectrum efficiency gain resulting from multiplexing different services on the same radio channel. This paper describes the key new concepts needed to support statistical multiplexing of different radio access bearer classes over the GSM/EDGE radio access network (GERAN) air interface. The focus is on the definition of new traffic and control channels to support statistical multiplexing of speech, real-time data, and non-real-time data, and the corresponding new MAC procedures that are needed to guarantee quality of service (QoS).

Cellular packet data systems require medium access control (MAC) protocols that allow flexible resource allocation on both the uplink and the downlink. The overhead introduced for managing MAC functions should be minimized, and control fields should be well protected in order to provide robustness to channel errors. The IS-136 TDMA digital control channel specifies shared channel feedback (SCF) procedures and fields for managing uplink resources. However, the SCF mechanism is not well suited to long packet data transactions. We describe a simple and robust packet channel feedback (PCF) mechanism for an IS-136 based TDMA packet data channel by identifying the minimal set of functions and fields required to manage both contention access and reserved access. The proposed scheme has unambiguous state transitions and better performance under error conditions. The PCF scheme provides considerable flexibility in assigning bandwidth to multiple users, and allows the efficient scheduling of contention and reservation opportunities on each subchannel without wasted slots.

Capacity analysis in CDMA systems is quite difficult due to the system complexity and interdependence of performance of all connections. Analytical models usually suffer from many simplifying approximations and presumed uniformity of traffic conditions. On the other hand, event driven simulators are time consuming both in the software development and numerical study phases, especially when a multi-cell environment is considered. In this paper we describe a model which combines some advantages of the two approaches. In particular it can provide results in time comparable with analytical models while taking into account interdependencies between the connection parameters the same way as in the simulation model. The model can be compared to capacity evaluation at a given time instant in a simulation model but with a statistical representation of the mobiles. Hence the name statistical snap-shot. In the paper we analyze the main features of the model. Numerical results obtained from the analytic model are found to be within 10% of the results obtained from a detailed simulation study.

We propose a technique to measure channel quality in terms of signal-to-interference plus noise ratio (SINR) for the transmission of signals over fading channels. The Euclidean distance (ED) metric, associated with the decoded information sequence or a suitable modification thereof, is used as a channel quality measure. Simulations show that the filtered or averaged metric is a reliable channel quality measure which remains consistent across different coded modulation schemes and at different mobile speeds. The average scaled ED metric can be mapped to the SINR per symbol. We propose the use of this SINR estimate for data rate adaptation, in addition to mobile assisted handoff (MAHO) and power control. We particularly focus on data rate adaptation and propose a set of coded modulation schemes which utilize the SINR estimate to adapt between modulations, thus improving the data throughput. Simulation results show that the proposed metric works well across the entire range of Dopplers to provide near-optimal rate adaptation to average SINR. This method of adaptation averages out short-term variations due to Rayleigh fading and adapts to the long-term effects such as shadowing. At low Dopplers, the metric can track Rayleigh fading and match the rate to a short-term average of the SINR, thus further increasing throughput.

The advantages of code division multiple access (CDMA) for cellular voice have become well known, and IS-95-based systems have now been widely deployed. Attention is now focused on higher data-rate packet services for cellular systems. Although many packet multiple access schemes have been studied over the years, researchers have often studied single cell performance and ignored reuse. Moreover, direct sequence spread spectrum (DSSS) has been considered unsuitable for high data-rate packet multiple access since spreading limits the permitted data rates, DSSS requires large overhead (preambles) for acquisition and requires closed-loop power control. In this paper, we describe a scheme for high data-rate packet service using CDMA that addresses all of the above problems and has been standardized in Revision B of IS-95. A low rate fundamental code channel is maintained that eliminates the need for long preamble and provides closed-loop power control. Reuse is managed by the infrastructure through a "burst-level" admission control based on load and interference-level measurements at the base stations and mobiles. We report on the feasibility of such a burst-mode packet data service for cellular CDMA networks. The focus is not only on the performance of high data-rate users, but also on the impact on voice users sharing the CDMA band. We propose a multitiered performance analysis methodology consisting of a mix of static simulations, dynamic simulations at different time scales, and analytic methods to address the various feasibility issues: impact on coverage; capacity; power control; and effectiveness of burst admission algorithms. Based on the current study, we can conclude that the proposed approach is well suited for third-generation wideband CDMA systems being considered for standardization throughout the world.

We study a simple admission control scheme for high speed packet data service over the reverse link of an IS-95 based CDMA air interface. A preliminary version of this scheme was originally proposed by Nanda et al. (see Proc. IEEE Globecom 96, 1996) and admits data users on a burst level. The performance and design tradeoffs are examined through simulations. The basic tradeoff is the coverage of high speed packet data service and its impact on outage experienced by the voice users. The results indicate that through limiting the region where a high speed data service is available, the performance impact on voice users can be made small. It is possible to extend the region of coverage of high speed data by using more complex schemes proposed by Nanda et al.

Cellular digital packet data (CDPD) systems borrow idle radio channels from Advanced Mobile Phone Standard (AMPS) cellular networks to send short bursts of packet data traffic. Unless AMPS channels are dedicated to CDPD, utilization of unused AMPS channels results in increased interference for voice. We quantify this increase in interference and show that in many systems the interference "slack" from the radio engineering in AMPS can be exploited to deploy CDPD on channels shared with voice. In particular, we find that this is true for 3-sector cells with 7-cell reuse, a deployment scheme used in many cellular systems.

CDMA wireless systems are believed to require smaller fade margins (than TDMA or FDMA systems) because of their ability to get into what is known as soft handoff. A mobile in soft handoff maintains simultaneous radio links with multiple base-stations which enables it to make use of the best quality leg most of the time. We present an analysis of fade margins for systems allowing soft handoffs (e.g. CDMA) and those where only hard handoffs are possible (e.g. FDMA, TDMA). Our analysis is based on a model for hard handoffs with hysteresis and connection delay, and validates the claim that systems permitting soft handoff indeed require smaller fade margins.

In a two-tier microcell/macrocell cellular system, to keep the handoff rate at an acceptable level, low mobility users (with speed /spl upsi/V/sub 0/) should undergo handoffs at macrocell boundaries. We propose a procedure by which the mobile determines user mobility from microcell sojourn times and uses it for micro-macro cell selection at call origination and handoff. The probability of erroneous assignment of a mobile to a microcell or macrocell is shown to be significantly lower than previous approaches.

The cellular radio link is characterized by deep fades leading to long error bursts (lasting hundreds of milliseconds). Data transmission over such links results in large packet error rates (in the range 10/sup -3/ to 10/sup -1/). We have designed a radio link protocol (RLP) to achieve high throughput on such links. The RLP is based on frequent, complete or partial feedback of the receiver state. Performance results for the US digital cellular TDMA standard show that the scheme can provide the equivalent of 9.6 kb/s service per full-rate TDMA channel above a carrier to interference ratio of 18 dB.>

Packet reservation multiple access (PRMA) provides spatially dispersed voice and data terminals wireless access to a base station over a common short-range radio channel. The author presents an equilibrium point analysis of the joint voice data system. The analysis leads to two coupled nonlinear equations. By a judicious change of variables, the author calculates the equilibrium of the joint system by finding the roots of a univariate equation. The author provides expressions for voice packet dropping probability, the distribution of data packet delays and the system throughput, at equilibrium. For the joint voice data system the design problem is two dimensional. Elementary catastrophe theory leads to a bifurcation of the design space into useful and congested regions. The useful design space is further restricted by target values of the performance criteria. The study of a nominal system verifies the design methodology. Moreover, it shows that PRMA gracefully accepts low bit rate data terminals with moderate data packet delays (few hundred milliseconds), while simultaneously satisfying a 1% speech packet dropping criterion.>

The authors consider the encoding of image subbands with a tree code that is asymptotically optimal for Gaussian sources and the mean squared error (MSE) distortion measure. They first prove that optimal encoding of ideally filtered subbands of a Gaussian image source achieves the rate distortion bound for the MSE distortion measure. The optimal rate and distortion allocation among the subbands is a by-product of this proof. A bound is derived which shows that subband coding is closer than full-band coding to the rate distortion bound for a finite length sequence. The tree codes are then applied to encode the image subbands, both nonadaptively and adaptively. Since the tree codes are stochastic and the search of the code tree is selective, a relatively few reproduction symbols may have an associated squared error a hundred times larger than the target for the subband. Correcting these symbols through a postcoding procedure improves the signal-to-noise ratio and visual quality significantly, with a marginal increase in total rate.

It is proved that optimal coding of image subbands achieves the ultimate rate-distortion performance for Gaussian sources with squared-error distortion. Optimal performance is achieved only asymptotically with infinite-size source sequences. The approach of the rate-distortion function of finite-size sequences to its limit is investigated and a bound is derived that shows a potentially closer approach for coding subbands that for the full band. A suboptimal realization of an optimal tree coding method is used to encode both the full-band image and its subbands at 1.0 bit per pixel. The subband tree-coded image is more than 2 dB better in SNR and its overall performance is nearly comparable to that of the best schemes known to date.>