HD computing is a symbolic representation system which performs various learning tasks in a highly-parallelizable and binary-centric way by drawing inspiration from concepts in human long-term memory. However, the current HD computing is ineffective in extracting high-level feature information for image data. In this paper, we present a neuro-symbolic approach called NSHD, which integrates CNNs and Hyperdimensional (HD) learning techniques to provide efficient learning with state-of-the-art quality. We devise the HD training procedure, which fully integrates knowledge from the deep learning model through a distillation process with optimized computation costs due to the integration. Our experimental results show that NSHD provides high energy efficiency as compared to CNN, e.g., up to 64% with comparable accuracy, and can outperform the learning quality when more computing resources are allowed. We also show the symbolic nature of the NSHD can make the learning humnan-interpretable by exploiting the property of HD computing.

The ultimate goal of network resource allocation for video teleconferencing is to optimize the Quality of Experience (QoE) of the video. We consider the IPPP video coding structure with macroblock intra refresh, which is widely used for video teleconferencing. Generally, the loss of a current frame causes error propagation to subsequent frames due to the video coding structure. Therefore, to optimize the QoE, a communication network needs to be able to accurately predict the consequence of each of its resource allocation decisions. We propose a QoE prediction scheme by considering QoE models that use the per-frame PSNR time series as the input, thus reducing the QoE prediction problem to a per-frame PSNR prediction problem. The QoE prediction scheme is jointly implemented by the video sender (orMCU) and the communication network. Simulation results show that the proposed per-frame PSNR prediction method is fairly accurate, with an average error well below 1dB.

This paper considers a downlink LTE system where a basestation (eNodeB) is serving video traffic (generated possibly due to video teleconferencing applications) to many users in a cell. The video traffic is separated into multiple sub-streams (logical channels) based on the coding structure or priority of video packets. The objective is to maximize video quality by servicing appropriate video sub-streams and users while taking into account the resource constrained wireless channel. It is shown that one can obtain significant gains in video quality by determining the quality of service (QoS) parameters on a per sub-stream basis, rather than in `user-based' approach where all the sub-streams are lumped into one single stream. Motivated by video teleconferencing applications, we provide simulation results for the case where video traffic is separated into logical channels based on a Hierarchical-P encoding structure. Furthermore, we demonstrate the gains in video quality that can be obtained by allocating resources across logical channels. With the exploding growth in mobile video traffic expected to happen in 4G systems, this study suggests potential benefits of a logical channel based approach as compared to conventional user-based schemes.

The overlay approach to dynamic spectrum access recently proposed in information theory allows both primary users (PUs) and secondary users (SUs) to simultaneously access the same spectrum with comparable power levels while ensuring no degradation to the performance of PUs. However, this approach is based on a number of idealized assumptions that are difficult to satisfy in practice, and existing efforts to address this issue fall outside physical layer processing. In this paper, we propose a number of physical layer mechanisms to make the overlay approach practical. In particular, we leverage the broadcast nature of the wireless medium and the latest breakthrough in full-duplex radios to resolve the synchronization problem and to get around the non-causal assumption while naturally offering delay diversity.

Summary form only given. We describe the design of a mobile streaming system, which optimizes video delivery based on dynamic analysis of user behavior and viewing conditions, including user proximity, viewing angle, and ambient illuminance.

The HSDPA mode of the 3GPP UMTS FDD standard extends the peak downlink rate of the W-CDMA waveform to 14 Mbps. To achieve high rates across a multipath channel the demodulator must include an advanced receiver rather than a conventional Rake. This paper presents the performance of an NLMS-based adaptive equalizer that supports mobile velocities as high as 250 km/h. The equalizer works with radio and implementation impairments typical of a non-HSDPA receiver, so redesign of the terminal front end is not required. A key aspect of the design is dynamically adjusting the equalizer step size and leakage parameter. There is often concern about the ability of an adaptive algorithm to operate in a highly mobile environment. The transient behavior is discussed in the context of HSDPA operation and examples are provided of initial convergence and operation after a sudden change in propagation conditions.

Signal-to-noise ratio (SNR) is an important parameter in many receivers. In a previous letter B Li et al., (2002) we derived a maximum-likelihood amplitude estimator and obtained a low bias SNR estimation algorithm for a binary phase-shift keying (BPSK) modulated signal. We now extend the result and obtain a low bias SNR estimator for multiple phase-shift keying (MPSK) modulation. We compare our algorithm with others, such as a decision-directed approach. Simulations show that the new algorithm provides not only much lower bias and but also much smaller mean square error (MSE).

This paper presents an efficient FFT based channel-equalization method for data detection. The data detection method is applicable to multiuser detection in the downlink, and single user detection in the uplink. It can operate for chip rate sampling as well as higher sampling rates. The FFT implementation is enabled by approximating the channel correlation matrix by a right circulant matrix. Furthermore, the circulant matrix approximation is optimized via a simple permutation procedure. The solution offers very low complexity relative to other receivers. For instance, for twelve codes with spreading factor 16 each, the complexity of the proposed solution is about 25% of the complexity of the approximate Cholesky based joint detector. The proposed FFT based solution significantly outperforms the matched filter (MF) and is comparable in performance to the approximate Cholesky based joint detector.

We consider the problem of detecting a subspace signal when there is uncertainty in the subspace. Such uncertainty usually causes a mismatch between the detector and the signal to be detected, which may lead to significant loss in performance. To improve the robustness of the detection procedure we apply robust adaptive subspace detectors based on extending the dimension of the signal subspace. We consider two types of adaptive constant false alarm rate (CFAR) detector structures for the extended subspace detectors: CFAR generalized likelihood ratio detector (CFAR GLR) and CFAR matched subspace detector (CFAR MSD). Using Monte-Carlo simulations, we study the performance of the robust adaptive subspace detectors for space time processing.

The problem of separating and estimating signals received by an array whose array manifold has an unknown structural form is usually referred to as the blind signal copy problem. In this paper we consider the blind signal copy problem for polynomial phase signals. By deriving the Cramer Rao bound we evaluate the optimal performance achievable by any unbiased estimator. To gain additional insight into this problem we compare the CRB to the bound for the case where the functional form of the array manifold is known. We derive a computationally efficient approximate maximum-likelihood (ML) algorithm and compare its performance with the bound.

Signal cancellation causes the performance of conventional adaptive beamformers to severely deteriorate in environments containing interference sources that are correlated with the signal. We propose an algorithm which overcomes the signal cancellation problem for wideband signals. The proposed algorithm applies array manifold interpolation to perform frequency domain averaging, which reduces the correlation between the signals and the interference sources. In contrast to other algorithms that apply frequency averaging, the algorithm proposed does not require any preliminary estimation or prior information on the interference environment.

This paper attempts to assess the potential performance gain of spatial-temporal processing relative to conventional spatial processing, for signals obeying a deterministic parametric model. The Cramer-Rao bound (CRB) on the estimates of the source directions of arrival (DOA) is used to quantify this gain. Spatial-temporal processing does not yield any such gain in the single source case, or for multiple coherent signals. However, significant gains can be achieved for multiple non-coherent signals.

Signal-to-noise ratio (SNR) thresholds in time delay estimation are discussed. At very low signal-to-noise ratios, mean square estimation errors are dominated by prior information because of the high probability of large errors in an extended search region. At higher SNR values mean square errors are dominated by the measured data unless the lack of prior information is extreme. Barankin bound techniques are used to show that the transition occurs in the range 1>

The accuracy of time delay estimates obtainable in active localization systems is investigated. The authors concentrate on narrowband signals for which, due to ambiguities in the time delay estimates, the Cramer-Rao bound is not achievable for signal-to-noise ratio (SNR) values below a certain threshold. The dual problems of determining the threshold SNR below which the Cramer-Rao is not achievable and obtaining a realizable bound for SNR values below that threshold is addressed. Thresholds and realizable bounds for the single and multiple echo problems are evaluated by applying the Barankin bound. It is shown that the thresholds are essentially independent of the target separation.>

The Cramer-Rao bound (CRB) for the time delays of closely spaced echoes is derived. It is shown that the variance of the ith echo delay estimate depends on the relative delays of all other echoes, but not on their amplitudes. This results from the restriction that the estimator must be unbiased. A single replica correlation algorithm is suggested for the case of two echoes with highly unequal amplitudes, and its performance is evaluated. For echo spacings below some threshold, the estimate for the dominant echo is slightly biased, but its mean square error is significantly lower than the CRB. The estimation of the weak echo is discussed.>