Seong-Hyun Jang

A new packet detection algorithm for IEEE 802.15.4a DBO-CSS in AWGN channel

In this paper, we propose a new packet detection algorithm for differentially bi-orthogonal chirp-spread-spectrum (DBO-CSS) over multiple piconets. Because of the band hopping property of the DBO-CSS, the conventional packet detection algorithms based on auto-correlation cannot avoid side band noise. And the conventional algorithms employed in DBO-CSS systems are not performed well over simultaneously operating piconets (SOP) because they did not consider the waveforms of signal. First, the proposed detector includes matched filters to eliminate the side band noise and to make SNR at the end of matched filters maximal. Second, the proposed detector also includes auto-correlators in order to make outputs of matched filters inphase. Finally, the proposed detector gathers the energy of all sub-chirps and employs the Permutator in order to operate regardless of piconet. Simulation results show that we can detect packets by using the proposed algorithm compared with the conventional one in about -4 dB SNR with the same miss detection probability and the false alarm probability between them.

RAKE Receiving Methods for IEEE 802.15.4a CSS System

In this paper, we propose two RAKE receiving methods of the chirp spread spectrum (CSS) system to suppress the influence of the multipath-fading channel. The first one has low complexity by using the simple accumulator to collect the energies of the multi-paths. The second one can enhance the PER performance of the differential detection by adopting the concept of the multiple-symbol differential detection (MS-DD). We verified the proposed architectures through Monte-carlo simulations in diffuse channels. The simulation results show that the RAKE receiving methods can efficiently reduce performance degradation in multipath channel.

Robust and Accurate Packet Detection and Frequency Offset Compensation for CSS System

In this paper, we propose a new robust packet detection and frequency offset compensation structure for IEEE 802.15.4a chirp spread spectrum (CSS) which is a standard for wireless personal area network (WPAN). The proposed structure takes benefits of both cross- and auto-correlation. Cross-correlation is useful to remove additional side band noise in the CSS system and to find sharp peaks among filtered output values. In addition, we use the auto-correlation in order to suppress the effect of frequency offset and the initial channel phase. Simulation results show that the proposed procedure is robust to the frequency offset, and can detect the presence of a signal in very low signal to noise ratio (SNR).

On the Capacity of MIMO-OFDM Systems with Doubly Correlated Channels

In this paper, the capacity of multiple input multiple output-orthogonal frequency division multiplexing (MIMO- OFDM) systems is analyzed under doubly correlated fading channels. The distribution of complex central Wishart Matrix is utilized to derive the closed form of ergodic capacity of MIMO- OFDM systems under doubly correlated channels when the transmitter has no channel state information (CSI) and the receiver has the perfect CSI. Then the obtaining results are used to study the impact of channel correlation on the capacity. The simulation results show that our closed form capacity is close to the real value and that the channel correlation degrades the system capacity.

Dual-band automatic gain control algorithm for chirp spread spectrum over indoor fading channels

In this paper, we propose a novel automatic gain control (AGC) algorithm as a solution to improve the signal reception performance based on the IEEE 802.15.4a chirp spread spectrum (CSS) over indoor fading channels. The CSS signals are composed of low-band and high-band signals whose frequency characteristics are different because of multipath fading in indoor environment. It is not appropriate to apply only one conventional AGC which regardless of the different bands, where the received signals are contaminated by quantization noise or clipped. The proposed AGC algorithm is based on two AGCs which are designed to be matched well with low-band and high-band signals separately.

Robust Packet Detection Algorithm for DBO-CSS

In this paper, we propose a robust packet detection algorithm for the differentially bi-orthogonal chirp-spread-spectrum (DBO-CSS). The conventional packet detection algorithms based on auto-correlation cannot avoid side band noise because of the band hopping property of the DBO-CSS. If we, therefore, use the conventional packet detection algorithms in the DBO-CSS, the performance degradation of the packet detector will break out. First, the proposed detector includes matched filters to eliminate the side band noise. Second, the proposed detector also includes auto-correlators in order to make outputs of matched filters inphase. Finally, the proposed detector gathers the energy of all sub-chirps by the summation. When we use the Permutator, the proposed detector works regardless of piconet. Simulation results show that we can detect packets in -4 dB SNR in AWGN channel and also in 0 dB in multipath channel.

Hybrid RSS/TOA wireless positioning with a mobile anchor in wireless sensor networks

In this paper, we propose a novel hybrid RSS/TOA positioning algorithm with a mobile anchor, which follows a spiral trajectory to cover the whole region of the wireless sensor networks (WSN). The conventional time of arrival (TOA) positioning algorithms have two considerable accuracy problems. Firstly, as the distance between the reference node and the sensor increases, its accuracy is degraded gradually. Secondly, when an obstacle exists between the reference node and the sensor, its accuracy is degraded greatly. In our proposed algorithm, the mobile anchor, which plays the same role as the reference node, moves on a spiral trajectory and transmits its signal periodically. The received signal strength (RSS) is utilized by the sensor not only to compute whether it is closed enough to the mobile anchor but also to distinguish the obstacle. As a result, the measurement range of the TOA ranging process is restricted within the interval of two adjacent spiral curves. Under the circumstance that no obstacle exists between the mobile anchor and the sensor, the TOA ranging process and the proposed two-step positioning calculation are operated to estimate the position of the sensor accurately. Simulations are performed to compare the performance of the proposed hybrid RSS/TOA positioning algorithm with the performance of the conventional TOA least square estimating positioning algorithm (TOA_LSE). Simulation results show that the average positioning error is reduced by 48.2% and the standard deviation of the positioning errors is reduced by 71.3%.

Two-step automatic gain control algorithm applicable to chirp spread spectrum device

In this paper, we propose a novel automatic gain control (AGC) algorithm, which is applicable to a commercially available chirp-spread-spectrum (CSS) device. Because of the property of long guard time in the CSS, it is difficult to make use of the conventional AGC algorithm which only uses one window size to estimate the power of signal. Due to using only one window size, there is a tradeoff between adaptation time and stability. The proposed AGC algorithm synthetically utilizes large window size and small window size, simultaneously improving the performance of adaptation time and stability.

Differential Super-Quasi-Orthogonal Space-Time Trellis Codes

This paper proposes a new differential super-quasi- orthogonal space-time trellis codes (SQOSTTC) with single- symbol maximum-likelihood (ML) decoding. Firstly, a new joint constellation, which can be used to construct a unitary matrix for quasi-orthogonal space-time block code (QOSTBC), is designed. Secondly, based on the joint constellation and criterion of maximizing the coding gain distance (CGD), set-partitioning is done. Finally, we introduce the simple single-symbol maximum- likelihood (ML) decoding. Simulation results show that our differential SQOSTTC can get better performance than the differential super-orthogonal space-time trellis code (SOSTTC) [5] with four transmit antennas at the transmission rate of 1.5 bits/s/Hz and lower decoding complexity than the differential super-pseudo-orthogonal space-time trellis codes (SPOSTTC) [13] and the differential SOSTTC [14] at the transmission rate of 2 bits/s/Hz.

Low complexity time delay estimation using propagator based on matrix partitioning

In this paper, low complexity time delay estimation (TDE) method based on data matrix partitioning is proposed. The dimension of signal subspace was increased by stacking data matrices for performance improvement of TDE. Increment of the size of data matrix causes high computational complexity of TDE. The proposed algorithm, which only requires propagator with smaller data matrix than original data matrix, can avoid high computational cost. Each of the partitioned matrices is used for the derivation of the averaged propagator matrix. The estimation error of the proposed method is approximately identical to that of the conventional propagator with less computational complexity by using the averaged propagator matrix.