We present a switching Kalman filter model for the real-time inference of hand kinematics from population motor cortical neurons. Firing rates are modeled as Gaussian mixture where mean each component is linear function kinematics. A "hidden state" models probability and evolves over time in Markov chain. The generalizes previous encoding decoding methods, addresses non-Gaussian nature firing rates, can cope with crudely sorted neural data common on-line prosthetic applications.

In this paper, an effective joint measurement selection and power allocation (JMSPA) scheme is proposed for the distributed multi-target tracking task in radar networks with missing data. Missing data may occur during exchange between radars a fusion center (FC) due to unreliability of communication channels. First, we derive predicted conditional Craméer-Rao lower bound (PC-CRLB) presence origin uncertainty, which are scaled by improved information reduction factor (IIRF). Second, overall...

This paper deals with the problem of detecting a moving range-spread target in distributed MIMO radar. A new knowledge-aided (KA) model that takes into account nonhomogenous characteristics disturbance (clutter and noise) radar is proposed. Specifically, covariance matrices corresponding to different transmit-receive (Tx-Rx) pairs are modeled as random matrices. These share prior matrix structure but power levels nonhomogeneous clutter powers across Tx-Rx pairs. Two cases considered,...

This paper addresses the problem of detecting a signal in partially homogeneous and environments. In environments, i.e., both test data training share same noise covariance matrix structure up to an unknown scaling factor, persymmetric adaptive coherence estimator (Per-ACE) detector is proposed. By exploiting matrix, Per-ACE can reduce requirements. Furthermore, expressions for probabilities false alarm detection are derived along with distribution loss factor β. matched filter (Per-AMF) has...

The problem of detecting a subspace signal in colored Gaussian noise with unknown covariance matrix is investigated by incorporating persymmetric structure received data. interest described model, namely, it belongs to spanned the columns known matrix, but coordinates. We propose detector two tunable parameters, which includes many existing detectors as special cases. Approximate expressions for probabilities false alarm and detection proposed are derived, verified via Monte Carlo...

We focus on the problem of detecting a signal in compound-Gaussian clutter, where texture is random variable with Gamma or inverse distribution. The persymmetric structure covariance matrix exploited and generalized likelihood ratio test (Per-GLRT) using three-step procedure proposed. In addition, we prove that Per-GLRT ensures constant false alarm rate (CFAR) property respect to matrix. Finally, detector assessed by Monte Carlo simulations. Performance comparison traditional GLRT shows...

Efficient motion parameter estimation is a key challenge for moving-target imaging and localization in the synthetic aperture radar ground indication system. However, existing methods suffer from ambiguities, complex realization, or heavy computation load of O(MN). To solve these problems, we propose an efficient Radon transform (RT) fractional Fourier (FRFT) to estimate radial velocity azimuth velocity. By exploiting geometry information, model relationship between parameters two angles RT...

This letter addresses the problem of distributed target detection in structured interference and non-homogeneous disturbance. The signal locate two linearly independent subspaces with unknown coordinates, while disturbance is partially homogeneous an covariance matrix. By incorporating persymmetric structure received data, we propose a subspace detector for target, which includes point-like as special case. Remarkably, proposed shown to ensure constant false alarm rate property respect both...

This letter addresses the problem of high-resolution radar detection in interference and nonhomogeneous noise. The target signal lie two linearly independent known subspaces, but with unknown coordinate. noise is modeled by a compound-Gaussian process covariance matrix random texture. According to two-step generalized likelihood ratio test-based design approach, we derive distributed detector. Numerical examples show that proposed detector can provide better performance than their...

We recently proposed a direction-of-arrival (DOA) estimation method for the coexistence of uncorrelated and coherent signals, which is related to uniform linear antenna arrays with mutual coupling (MC) in presence unknown nonuniform noise. This technique, however, relies on assumption mixed signals not suitable scenarios purely signals. paper extends signal scenario one hand, other allows us directly adopt ESPPRIT algorithm perform DOA MC environments. To be specific, help least squares (LS)...

For weak target detection and interference suppression, we consider a joint design of transmit waveform space-time receive filter to reduce the auto-correlation peak sidelobe level (APSL) cross-correlation (ICPL). First, propose criterion, which formulates multi-Lp norm signal-interference-noise-ratio (SINR) model under constraints mismatched suppress digital radio frequency memory (DRFM) interference. Second, present an iterative penalty majorizationminimization (IPMM) algorithm solve...

We consider a new hybrid radar paradigm consisting of an active array and passive array. In this system, the transmits probing signal from its receives two types echoes: One is return cooperative transmission other target due to noncooperative illuminators opportunity. The motivation for approach exploit not only but also any signals that are present in surveillance area, thereby maximizing signal-to-interference-plus-noise ratio (SINR). Numerical results demonstrate proposed concept can...

This paper deals with the problem of detecting a subspace signal in presence spatially and temporally colored disturbance. A new parametric model that takes into account multi-rank structure for target employs multi-channel auto-regressive process disturbance is proposed. Following this model, Rao detector (SP-Rao) developed training-limited scenarios. Unlike conventional detectors are designed only rank-one detection, SP-Rao has pairwise successive spatio-temporal whitening...

This study addresses the problem of detecting a distributed target in interference and noise with low sample support. The signal lie two linearly independent subspaces spectral property is unknown. number available training data too small to form non‐singular covariance matrix. To overcome difficulty, authors resort Bayesian framework design generalised likelihood ratio tests. Both detectors can reduce requirement by utilising prior knowledge Numerical results show that proposed provide...

In this study, the authors deal with problem of detecting a signal in partially homogeneous environments, where both test data and training share same covariance matrix up to an unknown scaling factor. A generalised persymmetric parametric adaptive coherence estimator (GPer‐PACE) detector is proposed, disturbance modelled as multichannel autoregressive process. To mitigate effect limited samples, subspatial aperture smoothing performed design authors’ GPer‐PACE detector. Moreover, structure...

Two-way opportunistic relaying (TWOR) systems in amplify-and-forward (AF) strategy have been widely studied because of simplicity and high spectral efficiency. However, there few investigations into the performance these systems. In this study, authors present a analysis equal power allocation (EPA) scheme for TWOR-AF over independent non-identically distributed Rayleigh fading channels. Closed-form lower upper bounds as well an approximation outage probability are established to show...

The perfectly partly calibrated antenna array is a frequently assumption in most of the existing gain/phase calibration methods. In practice, however, usually not available. this letter, tail optimization method for direction finding with unknown gains and phases presence spatially non-uniform noise proposed. Specifically, entry firstly merged into signal power by using sparse representation. Subsequently, that can significantly suppress occurrence pseudo-peaks designed to determine DOAs...

In this paper, a new method is proposed for direction-of-arrival (DOA) estimation of coherent signals with improved sparse representation in unknown spatially correlated Gaussian noise. To be specific, leveraging symmetric uniform linear array, the entries signal covariance matrix firstly recasted to eliminate Subsequently, it shown that an equivalent source vector can obtained by squaring any row noise-free matrix, irrespective coherency between signals. Finally, representation, which...

Multichannel adaptive signal detection jointly uses the test and training data to form an detector, then make a decision on whether target exists or not. Remarkably, resulting detectors usually possess constant false alarm rate (CFAR) properties, hence no additional CFAR processing is needed. Filtering not needed as procedure either, since function of filtering embedded in detector. Moreover, exhibits better performance than filtering-then-CFAR technique. has been more 30 years first...