Wenfei Guo

Improving the Design of MEMS INS-Aided PLLs for GNSS Carrier Phase Measurement under High Dynamics

The phase locked loop (PLL) bandwidth suffers a dilemma on carrier phase accuracy and dynamic stress tolerance in stand-alone global navigation satellite systems (GNSS) receivers. With inertial navigation system (INS) aiding, PLLs only need to tolerate aiding information error, instead of dynamic stress. To obtain accurate carrier phase under high dynamics, INS-aided PLLs need be optimally designed to reduce the impact of aiding information error. Typical micro-electro-mechanical systems (MEMS) INS-aided PLLs are implemented and tested under high dynamics. Tests using simulation show there is a step change in the aiding information at each integer second, which deteriorates the carrier phase accuracy. An improved structure of INS-aided PLLs is proposed to eliminate the step change impact. Even when the jerk is 2000 m/s3, the tracking error of the proposed INS-aided PLL is no more than 3°. Finally, the performances of stand-alone PLLs and INS-aided PLLs are compared using field tests. When the antenna jerk is 300 m/s3, the carrier phase error from the stand-alone PLLs significantly increased, while the carrier phase error from the MEMS INS-aided PLLs almost remained the same. Therefore, the proposed INS-aided PLLs can suppress tracking errors caused by noise and dynamic stress simultaneously under high dynamics.

Low-end MEMS IMU can contribute in GPS/INS deep integration

In a deeply-coupled GPS/INS integrated system, the use of the inertial aiding information can improve the tracking loop performance and make the system more robust. To meet this requirement, the inertial aiding information should have sufficient accuracy in short-term (such as the sampling interval of GPS, e.g. 1sec). The MEMS (Micro-Electro Mechanical System) IMU (Inertial Measurement Unit) can be a promising candidate due to its small size and low cost. There should be no doubt that MEMS INS (Inertial Navigation System) can aid the GPS receiver tracking loop by eliminating the dominant part of the motion dynamic stress, considering that the INS errors induced by the receiver motion dynamics is much less than the motion dynamic itself, when the receiver manoeuvres. So the only concern the side effect caused by MEMS INS, which determine whether MEMS IMU is qualified for deep integration, is its navigation error independent with the motion dynamics (i.e. manoeuvre-independent error). This paper assesses this side effect of MEMS INS in terms of providing Doppler aiding data in to the GPS carrier tracking loop through a thorough error propagation analysis. The Laplace transform analysis is applied to the simplified INS error dynamic equations under stationary condition and find out the transfer relation between the error sources and the velocity estimation errors. Then the velocity error is converted to Doppler aiding error and substitute into the GPS tracking loop to analyze the corresponding carrier phase error. Results show that the largest velocity error caused by maneuver-independent errors is less than 0.1m/s during the typical GPS update interval (e.g. 1 sec), which meets the real road test results. The consequent carrier phase tracking error caused by the maneuver-independent error of MEMS INS is below 1.2 degree, which is much less than receiver inherent errors (e.g. the oscillator error and thermal noise). Conclusion can be reached that even the low-end MEMS IMUs have the ability of aiding the GPS receiver signal tracking although it induces some additional errors.

Double Stage NCO-Based Carrier Tracking Loop in GNSS Receivers for City Environmental Applications

A double stage NCO (numerically controlled oscillator)-based tracking loop with a new frequency discriminator is proposed to track the carrier frequency of the GNSS signal in signal degraded environments. The structure can be easily implemented and modified with conventional GNSS receivers, and the new frequency discriminator has absolute linearity in a wide pull in range. The principle and performance of the method are described and analyzed in detail. Simulation and field test with a software receiver show that this method can effectively track a weak signal compared with other presented methods in an open loop.

How the Integral Operations in INS Algorithms Overshadow the Contributions of IMU Signal Denoising Using Low-Pass Filters

This paper has made a comprehensive investigation of the contribution of inertial measurement unit (IMU) signal denoising in terms of navigation accuracy, through theoretical analysis, simulations and real tests. Analysis shows that the integral step in the inertial navigation system (INS) algorithm is essentially equivalent to a super low-pass filter (LPF), whose filtering strength is related to the integral time of the INS. Therefore the contribution of the IMU denoising filter is almost completely overshadowed by the effect of the integral step for normal navigation cases. The theoretical analysis result was further verified by the simulations with an example of inertial angle estimation and by real tests of INS and GPS/INS systems. Results showed that the IMU signal denoising cannot bring observable improvement to INS or GPS/INS systems. This conclusion is strictly valid in the condition that the equivalent cut-off frequency of the integral step (which equals the reciprocal of the INS working alone time) is lower than the cut-off frequency of the denoising filter, which is the usual case for INS applications (except for some static data processing such as the stationary alignment of INS).

Correlation acceleration in GNSS software receivers using a CUDA-enabled GPU

The correlation process in a GNSS receiver tracking module can be computationally prohibitive if it is executed on a central processing unit (CPU) using single-instruction single-data algorithms. An efficient replacement for a CPU is a graphics processing unit (GPU). A GPU is composed of massive parallel processors with high floating point performance and memory bandwidth. It can be used to accelerate the burdensome correlation process in GNSS software receivers. We propose a novel GPU-based correlator architecture for GNSS software receivers, which is independent of the GPU device, the number of the processing channels, the signal type, and the correlation time. The proposed architecture is implemented and optimized using CUDA, a parallel computing platform and programming model for GPUs. We focus on the following aspects: the design and the time complexity analysis of the proposed GPU-based correlator algorithm, the tests that verify the correctness and the optimization of the implementation, and the performance evaluation of the optimized GPU-based correlator. Moreover, we introduce some new CUDA features that can be applied in a GPU-based correlator.

Parameters Design Method of Kalman Filter-Based Tracking Loop in GNSS/INS Deep Integration

This paper proposes a parameter design method of a second-order KF tracking loop in the deep integration architecture according to inertial sensor errors and receiver oscillator errors. Firstly, the effects of INS aiding on the state-driven and measurement noises of KF tracking loop are analyzed. Secondly, through coordinate frame transformation and LOS projection, the variances of the two accelerometer noises in the Earth-centered Earth-fixed (ECEF) frame and LOS direction are obtained. Lastly, the performance of KF-based tracking loop in the deep integration architecture is tested through the data collected from the GNSS/INS hardware simulator, and is compared with the conventional tracking loop in the deep integration architecture. The test results show that the proposed parameter design method makes the channel KF in the deep integration share the comparable performance to the conventional tracking loop in the deep integration. The test results illustrate the correctness and effectiveness of the parameter design method proposed in the paper. Since each parameter is set according to its physical meaning in the paper, the empirical and tentative trials in the parameter optimization process are avoided.

Analytical and simulation-based comparison between traditional and Kalman filter-based phase-locked loops

We investigate and quantitatively analyze the similarities and differences between traditional phase-locked loops (PLLs) and Kalman filter (KF)-based PLLs. We focus on three aspects. First, the transient response of traditional and KF-based PLLs versus the steady-state bandwidth is investigated, where the transient time is used as a criterion to compare their tracking performance in the transient state. Second, the noise performance and the dynamic response of the PLL are investigated and compared, in the cases that the Doppler shift is extracted from either the velocity accumulator or the filter output. Moreover, the steady-state frequency error due to the high signal dynamics is derived. Third, a method that feeds back the frequency rate to numerically controlled oscillator (NCO) is proposed, where a detailed derivation of the quantitative mathematical relationship between the phase mismatch, the coherent integration interval and the signal dynamics is performed. The proposed method provides an effective approach to reducing the significant phase mismatch and obtaining more accurate carrier phase measurements in applications where high dynamics or long integration intervals may influence the tracking performance of the GNSS receivers.

Design and Performance Evaluation of a Dual Antenna Joint Carrier Tracking Loop

In order to track the carrier phases of Global Navigation Satellite Systems (GNSS) signals in signal degraded environments, a dual antenna joint carrier tracking loop is proposed and evaluated. This proposed tracking loop processes inputs from two antennas, namely the master antenna and the slave antenna. The master antenna captures signals in open-sky environments, while the slave antenna capture signals in degraded environments. In this architecture, a Phase Lock Loop (PLL) is adopted as a master loop to track the carrier phase of the open-sky signals. The Doppler frequency estimated by this master loop is utilized to assist weak carrier tracking in the slave loop. As both antennas experience similar signal dynamics due to satellite motion and clock frequency variations, a much narrower loop bandwidth and possibly a longer coherent integration can be adopted to track the weak signals in slave channels, by utilizing the Doppler aid from master channels. PLL tracking performance is affected by the satellite/user dynamics, clock instability, and thermal noise. In this paper, their impacts on the proposed phase tracking loop are analyzed and verified by both simulation and field data. Theoretical analysis and experimental results show that the proposed loop structure can track degraded signals (i.e., 18 dB-Hz) with a very narrow loop bandwidth (i.e., 0.5 Hz) and a TCXO clock.

A New SIMD Correlator Algorithm for GNSS Software Receivers to Process Complex IF Data

GNSS software receivers implement digital signal processing algorithms on programmable software platform (such as PC, DSP), which traditional ones execute on dedicated hardware. It is highly flexible, convenient for debug and could be adapted to a platform for GNSS algorithm research. However, the computation of down shifting of the GNSS signal to baseband and correlation with the locally generated ranging code is too expensive for normal GNSS software receivers to work in real time. The computational cost could be reduced by utilizing Single Instruction Multiple Data (SIMD) operations. The article proposes a new SIMD correlator algorithm for complex GNSS IF signal processing on x86 processors. It firstly demonstrates why normal GNSS software receivers couldn’t achieve real-time processing by using Single Instruction Single Data (SISD) operations and the improvements achieved by using existing SIMD algorithms; then proposes a new SIMD correlator algorithm, outlines its implementation principle and compares it with the SISD and the existing SIMD algorithms. Performance gains achieved via the new SIMD algorithms are then demonstrated in an analysis. Finally, the implementation and experimental results of the new algorithm are presented. The experimental results show that compared to the SISD and existing SIMD algorithms the new SIMD algorithm can effectively reduce the computation of down shifting of the GNSS signal to baseband and correlation with the locally generated ranging code.