Markus Leinonen

Distributed Joint Resource and Routing Optimization in Wireless Sensor Networks via Alternating Direction Method of Multipliers

We consider a distributed total transmit power minimization in a multi-hop single-sink data gathering wireless sensor network by jointly optimizing the resource allocation and the routing with given source rates. An inherent coupling in optimal routing and resource allocation is taken into account via cross-layer optimization to increase the energy efficiency of the network. Instead of distributing the solution process horizontally by commonly used dual decomposition, we apply consensus optimization in conjunction with the alternating direction method of multipliers (ADMM). By duplicating flow variables, the problem decomposes into node specific subproblems with local variables. These variables are iteratively driven into consensus via the ADMM. Numerical examples show that the proposed algorithm converges significantly faster as compared to the state of the art methods based on the dual decomposition. Additionally, the algorithm is appealing for practical implementation due to its low local communication overhead, robust operation in slightly changing channel conditions and scalability to large networks.

Distributed correlated data gathering in wireless sensor networks via compressed sensing

This paper considers energy efficient distributed data gathering methods for correlated data field monitoring in multi-hop wireless sensor networks (WSNs) via compressed sensing (CS). This signal compression and acquisition method can be used to exploit the inherent temporal and spatial correlation of the sensor signals. Many existing CS-based data gathering methods exploit correlation structures only in one signal dimension in WSNs. Thus, we propose a distributed method which utilizes the joint correlation pattern of a multi-dimensional WSN signal ensemble via Kronecker sparsity basis. Numerical results how that the proposed method can significantly reduce the data traffic in correlated data gathering multi-hop WSNs.

Compressed acquisition and progressive reconstruction of multi-dimensional correlated data in wireless sensor networks

This paper addresses compressed acquisition and progressive reconstruction of spatially and temporally correlated signals in wireless sensor networks (WSNs) via compressed sensing (CS). We propose a novel method based on sliding window processing, where the sink periodically collects CS measurements of sensor samples, and then, instantaneously reconstructs current WSN samples by exploiting the spatio-temporal correlation via Kronecker sparsifying bases. By using previous estimates as prior information, the method can progressively improve the reconstruction accuracy of the signal ensemble. Furthermore, the method can control the trade-off between decoding delay and complexity. Numerical results demonstrate that the proposed method can recover WSN data samples from CS measurements with higher reconstruction accuracy, yet with lower decoding delay and complexity, as compared to the state of the art methods.

Distributed consensus based joint resource and routing optimization in wireless sensor networks

We consider distributed total transmit power minimization in single-sink data gathering wireless sensor network via joint optimization of routing and resource allocation with given source rates. There exists numerous decentralized cross-layer optimization algorithms derived via dual decomposition over the physical and network layer. We propose a consensus based distributed optimization algorithm which applies alternating direction method of multipliers to efficiently solve the problem with low message passing. Numerical examples illustrate how significantly faster the proposed algorithm converges as compared to the state of the art method based on dual decomposition.

Consensus based distributed joint power and routing optimization in wireless sensor networks

This paper proposes a fast distributed optimization algorithm for total transmit power minimization in single-sink data gathering wireless sensor networks. Many of the existing decentralized optimization algorithms addressing cross-layer design over the physical and network layer are based on dual decomposition. Our design includes joint power and routing optimization with given source rates by using consensus mechanism in conjunction with alternating direction method of multipliers (ADMM). Thus, the problem is decoupled across the nodes via introducing local copies of the variables, which are then iteratively driven into consensus with the ADMM. By the numerical experiments, the proposed distributed algorithm is shown to converge significantly faster to near optimal solutions with a small amount of local variable exchange as compared to the existing methods based on the dual decomposition.

Channel-robust compressed sensing via vector pre-quantization in wireless sensor networks

This paper addresses channel-robust compressed sensing (CS) acquisition of sparse sources under complexity-constrained encoding over noisy channels in wireless sensor networks. We propose a single-sensor joint source-channel coding method based on channel-optimized vector quantization by designing a CS-aware encoder-decoder pair to minimize the end-to-end mean square error (MSE) distortion of the signal reconstruction. As our key target is to obtain tolerable encoding complexity at the resource-limited sensor, the method relies on vector pre-quantization of the measurement space. We derive the necessary optimality conditions for the system blocks using alternating optimization. Numerical results show that our proposed method achieves higher robustness against the joint effect of CS reconstruction, quantization, and channel errors with lower encoding complexity as compared to state of the art CS methods.

Distributed Distortion-Rate Optimized Compressed Sensing in Wireless Sensor Networks

This paper addresses lossy distributed source coding for acquiring correlated sparse sources via compressed sensing (CS) in wireless sensor networks. Noisy CS measurements are separately encoded at a finite rate by each sensor, followed by the joint reconstruction of the sources at the decoder. We develop a novel complexity-constrained distributed variable-rate quantized CS method, which minimizes a weighted sum between the mean square error signal reconstruction distortion and the average encoding rate. The encoding complexity of each sensor is restrained by pre-quantizing the encoder input, i.e., the CS measurements, via vector quantization. Following the entropy-constrained design, each encoder is modeled as a quantizer followed by a lossless entropy encoder, and variable-rate coding is incorporated via rate measures of an entropy bound. For a two-sensor system, necessary optimality conditions are derived, practical training algorithms are proposed, and complexity analysis is provided. Numerical results show that the proposed method achieves superior compression performance as compared with baseline methods, and lends itself to versatile setups with different performance requirements.

Distributed variable-rate quantized compressed sensing in wireless sensor networks

This paper addresses distributed finite-rate quantized compressed sensing (QCS) acquisition of correlated sparse sources in wireless sensor networks. We propose a distributed variable-rate QCS compression method with complexity-constrained encoding to minimize a weighted sum of the mean square error distortion of the signal reconstruction and the average encoding rate. The variable-rate coding is realized via entropy-constrained vector quantization, whereas the restrained encoding complexity is obtained via vector pre-quantization of CS measurements. We derive necessary optimality conditions for the system blocks for two-sensor case. Numerical results show that our proposed method efficiently exploits the signal correlation, and achieves superior distortion-rate compression performance.

Rate-distortion lower bound for compressed sensing via conditional remote source coding

Lossy compressed sensing (CS) of a sparse source is studied. A lower bound to the best achievable compression performance in a finite rate CS setup is established by providing support side information to the encoder and decoder. The rate-distortion problem is formulated via remote source coding and conditional rate-distortion theory. The best encoder separates into an estimation step and a rate-dependent transmission step. Numerical results illustrate the rate-distortion behavior of the scheme.