Ji-An Luo
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Ji-An Luo has a total of 41 co-authors. Here are the top co-authors by number of shared publications:
- Dongliang Peng
School of Automation, Hangzhou Dianzi University, Hangzhou 310018, China
8 shared publications - Zhi Wang
Zhejiang University
6 shared publications - Xiao-Ping Zhang
Tsinghua University
4 shared publications Dong-Liang Peng
Hangzhou Dianzi University
3 shared publications- Han Shen-Tu
Institution of Information and Control Hangzhou Dianzi University Hangzhou China
3 shared publications - Ming Yin
Department of Control Science and Engineering, Zhejiang University, Hangzhou 310027, China
2 shared publications
Latest Publications
- 29 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 75 (00189456) , 6510918
On-Device Generalizable Acoustic NLoS Identification via IQ Trajectory Representation
With the rapid growth of Internet of Things (IoT) technologies, acoustic signal–based localization has been increasingly adopted. However, non-line-of-sight (NLoS) propagation introduces severe multipath and reverberation, motivating reliable NLoS identification for robust indoor deployment. This paper proposes a lightweight end-to-end LoS/NLoS classification framework. The received chirp is self-mixed to form an intermediate-frequency (IF) signal and then quadrature-demodulated to baseband. The resulting IQ components are converted into a two-dimensional IQ-trajectory image. An adaptive elliptic low-pass filter with a dynamically estimated cutoff frequency is introduced to mitigate reverberation. To improve cross-environment robustness, we employ two-view knowledge distillation, where the teacher uses weak augmentation and the student uses strong augmentation with a consistency objective. Experiments in three indoor scenarios achieve accuracies of 0.977, 0.986, and 0.978, and robustness is further validated under distance variation, small displacement, and additive noise. The INT8-quantized TinyCNN is ∼50 kB and runs in 6.8 ms on a Raspberry Pi 4 and 5.0–5.8 ms on Android. Code, data, and weights: https://github.com/ChirsJia/iq-nlos.
- 29 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 75 (00189456) , 6510918
On-Device Generalizable Acoustic NLoS Identification via IQ Trajectory Representation
With the rapid growth of Internet of Things (IoT) technologies, acoustic signal–based localization has been increasingly adopted. However, non-line-of-sight (NLoS) propagation introduces severe multipath and reverberation, motivating reliable NLoS identification for robust indoor deployment. This paper proposes a lightweight end-to-end LoS/NLoS classification framework. The received chirp is self-mixed to form an intermediate-frequency (IF) signal and then quadrature-demodulated to baseband. The resulting IQ components are converted into a two-dimensional IQ-trajectory image. An adaptive elliptic low-pass filter with a dynamically estimated cutoff frequency is introduced to mitigate reverberation. To improve cross-environment robustness, we employ two-view knowledge distillation, where the teacher uses weak augmentation and the student uses strong augmentation with a consistency objective. Experiments in three indoor scenarios achieve accuracies of 0.977, 0.986, and 0.978, and robustness is further validated under distance variation, small displacement, and additive noise. The INT8-quantized TinyCNN is ∼50 kB and runs in 6.8 ms on a Raspberry Pi 4 and 5.0–5.8 ms on Android. Code, data, and weights: https://github.com/ChirsJia/iq-nlos.
- 10 April 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Aerospace and Electronic Systems
- Vol. 61 (4) , 9879-9899
Exact and Approximate Solutions for Squared-Range Target Motion Analysis
This paper focuses on the squared-range target motion analysis (SR-TMA) when only range measurements are available. By squaring the range measurements, the motion parameters can be estimated using the constrained weighted leastsquares (WLS) approach by exploring the relationship between the motion vector and three extra variables. However, the SRTMA suffers from the problem of matrix ill-conditioning, which would inevitably degrade the estimation performance significantly. Additionally, the nonlinear constraints make the design of analytical solutions intractable. In this paper, we employ the variable separation (VS) technique to SR-TMA that yields computationally efficient algebraic solution without having the matrix ill-conditioned issue. Due to different ways of handling the nonlinear constraints between the extra variables and the motion parameters, we propose two solutions for SR-TMA. One utilizes the Taylor series expansion of nonlinear constraint functions to refine the unconstrained VSSR-WLS estimate, which is computationally attractive. The other is to seek exact solutions by decoupling extra variables from motion parameters in the constraint conditions, resulting in higher estimation accuracy. The constrained WLS is demonstrated theoretically to achieve the Cramer-Rao lower bound. Simulations validate the ´ performance of two proposed estimators.
- 10 April 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Aerospace and Electronic Systems
- Vol. 61 (4) , 9879-9899
Exact and Approximate Solutions for Squared-Range Target Motion Analysis
This paper focuses on the squared-range target motion analysis (SR-TMA) when only range measurements are available. By squaring the range measurements, the motion parameters can be estimated using the constrained weighted leastsquares (WLS) approach by exploring the relationship between the motion vector and three extra variables. However, the SRTMA suffers from the problem of matrix ill-conditioning, which would inevitably degrade the estimation performance significantly. Additionally, the nonlinear constraints make the design of analytical solutions intractable. In this paper, we employ the variable separation (VS) technique to SR-TMA that yields computationally efficient algebraic solution without having the matrix ill-conditioned issue. Due to different ways of handling the nonlinear constraints between the extra variables and the motion parameters, we propose two solutions for SR-TMA. One utilizes the Taylor series expansion of nonlinear constraint functions to refine the unconstrained VSSR-WLS estimate, which is computationally attractive. The other is to seek exact solutions by decoupling extra variables from motion parameters in the constraint conditions, resulting in higher estimation accuracy. The constrained WLS is demonstrated theoretically to achieve the Cramer-Rao lower bound. Simulations validate the ´ performance of two proposed estimators.
- 24 February 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Aerospace and Electronic Systems
- Vol. 61 (4) , 10839-10846
Modified Polar Localization in Wireless Sensor Array Networks
The Direct Position Determination (DPD) problem is studied by employing Modified Polar Representation (MPR) in the context of wireless sensor arrays. MPR serves to alleviate the range threshold effect detected by the Cartesian-based Hybrid Bhattacharyya-Barankin (HBB) bound. Nevertheless, the conventional MPR-based two-step method encounters performance degradation in low signal-to-noise ratio regions. To address this challenge, an MPR-based DPD framework is introduced. Firstly, we propose a MPR-based maximum likelihood estimator employing Gauss-Newton iteration. Secondly, the stability of the estimator is affirmed through its MPR-based HBB bound. Simulation results demonstrate that the proposed method delivers precise bearing estimates, irrespective of the source's proximity.
- 24 February 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Aerospace and Electronic Systems
- Vol. 61 (4) , 10839-10846
Modified Polar Localization in Wireless Sensor Array Networks
The Direct Position Determination (DPD) problem is studied by employing Modified Polar Representation (MPR) in the context of wireless sensor arrays. MPR serves to alleviate the range threshold effect detected by the Cartesian-based Hybrid Bhattacharyya-Barankin (HBB) bound. Nevertheless, the conventional MPR-based two-step method encounters performance degradation in low signal-to-noise ratio regions. To address this challenge, an MPR-based DPD framework is introduced. Firstly, we propose a MPR-based maximum likelihood estimator employing Gauss-Newton iteration. Secondly, the stability of the estimator is affirmed through its MPR-based HBB bound. Simulation results demonstrate that the proposed method delivers precise bearing estimates, irrespective of the source's proximity.
- 29 January 2025
- Published by Elsevier in Digital Signal Processing
- Vol. 160
A novel regularized least-squares estimator for range/bearings-only maneuvering target tracking with unknown noise statistics
- 29 January 2025
- Published by Elsevier in Digital Signal Processing
- Vol. 160
A novel regularized least-squares estimator for range/bearings-only maneuvering target tracking with unknown noise statistics
- Open Access
- 21 February 2024
- Published by Institution of Engineering and Technology (IET) in Electronics Letters
- Vol. 60 (4)
PCRLB for batch discrete‐time estimation through state decoupling
This letter aims to obtain a closed‐form performance metric using the smoothing posterior Cramér–Rao lower bound (PCRLB) for a class of direct discrete‐time kinematic models. By decoupling the state‐space model into separate position, velocity and acceleration components using an independent dynamic representation, the velocity component in the a priori distribution is converted into a position component. Based on the above derivations, the closed‐form PCRLB is developed. The effectiveness of the proposed bound is demonstrated by contrasting the proposed smoothing PCRLB with the traditional recursive PCRLB in a bearings‐only target tracking scenario.
- 21 February 2024
- Published by Institution of Engineering and Technology (IET) in Electronics Letters
- Vol. 60 (4)
Open AccessPCRLB for batch discrete‐time estimation through state decoupling
This letter aims to obtain a closed‐form performance metric using the smoothing posterior Cramér–Rao lower bound (PCRLB) for a class of direct discrete‐time kinematic models. By decoupling the state‐space model into separate position, velocity and acceleration components using an independent dynamic representation, the velocity component in the a priori distribution is converted into a position component. Based on the above derivations, the closed‐form PCRLB is developed. The effectiveness of the proposed bound is demonstrated by contrasting the proposed smoothing PCRLB with the traditional recursive PCRLB in a bearings‐only target tracking scenario.
- Open Access
- 10 December 2023
- Published by Institution of Engineering and Technology (IET) in IET Radar, Sonar & Navigation
- Vol. 18 (5) , 799-811
A novel efficient estimator for three‐dimensional bearings‐only source localisation with unknown sensor altitude and systematic measurement errors
A novel two‐stage profile maximum likelihood estimator is proposed to estimate the source location and the systematic errors jointly, with the aim of addressing the problem of source localisation using angle‐only measurements from a single sensor with unknown sensor altitude and systematic measurement errors. The proposed two‐stage profile maximum likelihood estimator algorithm is capable of decoupling the azimuth and elevation angle measurements while transforming the original maximum likelihood optimisation problem into two sub‐problems, that is, two‐dimensional‐projected maximum likelihood estimator and relative altitude maximum likelihood estimator. In terms of the two‐dimensional‐projected maximum likelihood estimator, an algorithm combining pseudo‐linear estimating and Kalman filtering is proposed to generate an initial estimate. Subsequently, a Gauss–Newton iterative method is developed to estimate the two‐dimensional‐projected target location and the azimuth systematic error jointly. The relative height maximum likelihood estimator is initialised using a pseudo‐linear estimator. Next, a Gauss–Newton iterative algorithm is adopted to estimate the elevation systematic error and the relative height. As indicated by the result of the simulation studies, the proposed algorithm exhibits an estimation performance close to the Cramér–Rao lower bound with unknown sensor altitude and systematic measurement errors.
- 10 December 2023
- Published by Institution of Engineering and Technology (IET) in IET Radar, Sonar & Navigation
- Vol. 18 (5) , 799-811
Open AccessA novel efficient estimator for three‐dimensional bearings‐only source localisation with unknown sensor altitude and systematic measurement errors
A novel two‐stage profile maximum likelihood estimator is proposed to estimate the source location and the systematic errors jointly, with the aim of addressing the problem of source localisation using angle‐only measurements from a single sensor with unknown sensor altitude and systematic measurement errors. The proposed two‐stage profile maximum likelihood estimator algorithm is capable of decoupling the azimuth and elevation angle measurements while transforming the original maximum likelihood optimisation problem into two sub‐problems, that is, two‐dimensional‐projected maximum likelihood estimator and relative altitude maximum likelihood estimator. In terms of the two‐dimensional‐projected maximum likelihood estimator, an algorithm combining pseudo‐linear estimating and Kalman filtering is proposed to generate an initial estimate. Subsequently, a Gauss–Newton iterative method is developed to estimate the two‐dimensional‐projected target location and the azimuth systematic error jointly. The relative height maximum likelihood estimator is initialised using a pseudo‐linear estimator. Next, a Gauss–Newton iterative algorithm is adopted to estimate the elevation systematic error and the relative height. As indicated by the result of the simulation studies, the proposed algorithm exhibits an estimation performance close to the Cramér–Rao lower bound with unknown sensor altitude and systematic measurement errors.
- 7 August 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Vehicular Technology
- Vol. 73 (1) , 28-44
Towards Fast and Accurate 3D Direct Emitter Tracking via Laplace Approximation
In this paper, the location and velocity of an emitter are estimated through spatially distributed receivers based on delay and Doppler-shift induced signal waveforms. In the context of Bayesian filtering, the main challenge in direct emitter tracking is to estimate the high-dimensional parameters based on the non-linear and non-Gaussian marginal posterior distribution. Although Monte Carlo approximations can be set as close to the optimal solution as expected, these methods usually incur a high computational cost and are inefficient in dealing with high-dimensional problems. Therefore, a novel Posterior Laplace Approximated Filter (PLAF) algorithm intended for direct emitter tracking is proposed in this paper to achieve a comparable estimation performance close to the posterior Cramér-Rao lower bound but at a low level of computational complexity. The PLAF method relies on the Laplace approximation technique to approximate the marginal posterior as a Gaussian distribution, the mean and covariance matrix of which are explicitly calculated. Furthermore, an improved approach to unconstrained optimization is developed based on a feature space search technique. The proposed optimization method is capable to jump out of the neighborhood of the local optimum to find a better one. The PLAF algorithm performs well in computationally efficiency because it belongs to a deterministic approximation domain. Besides, it can maintain a superior performance, even in the context of a low signal-to-noise ratio. Simulation results are obtained to demonstrate the performance of the proposed algorithm.
- 7 August 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Vehicular Technology
- Vol. 73 (1) , 28-44
Towards Fast and Accurate 3D Direct Emitter Tracking via Laplace Approximation
In this paper, the location and velocity of an emitter are estimated through spatially distributed receivers based on delay and Doppler-shift induced signal waveforms. In the context of Bayesian filtering, the main challenge in direct emitter tracking is to estimate the high-dimensional parameters based on the non-linear and non-Gaussian marginal posterior distribution. Although Monte Carlo approximations can be set as close to the optimal solution as expected, these methods usually incur a high computational cost and are inefficient in dealing with high-dimensional problems. Therefore, a novel Posterior Laplace Approximated Filter (PLAF) algorithm intended for direct emitter tracking is proposed in this paper to achieve a comparable estimation performance close to the posterior Cramér-Rao lower bound but at a low level of computational complexity. The PLAF method relies on the Laplace approximation technique to approximate the marginal posterior as a Gaussian distribution, the mean and covariance matrix of which are explicitly calculated. Furthermore, an improved approach to unconstrained optimization is developed based on a feature space search technique. The proposed optimization method is capable to jump out of the neighborhood of the local optimum to find a better one. The PLAF algorithm performs well in computationally efficiency because it belongs to a deterministic approximation domain. Besides, it can maintain a superior performance, even in the context of a low signal-to-noise ratio. Simulation results are obtained to demonstrate the performance of the proposed algorithm.
- 13 June 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Sensors Journal
- Vol. 23 (15) , 17078-17091
A Novel Range-Selective Generalized Weighted Centroid Method for Source Localization in Bearing-Only Sensor Networks
In this study, the source localization problem is investigated using a group of networked sensors if only bearing measurements are available. The existing bearing-only source localization methods are subjected to a critical problem, i.e., how to achieve a fine balance between the computational complexity and the positioning accuracy when the threshold effect is considered. A novel range-selective generalized weighted centroid localization (RSGWCL) estimator is proposed to address the above-mentioned problem. Without the threshold effect, the proposed RSGWCL method is capable of achieving comparable positioning accuracy to the maximum likelihood estimator (MLE). If the threshold effect is identified, the RSGWCL estimator still has a capability of providing reliable estimates and outperforms the MLE. The result of the simulation is analyzed, such that the performance of the proposed method can be verified.
- 13 June 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Sensors Journal
- Vol. 23 (15) , 17078-17091
A Novel Range-Selective Generalized Weighted Centroid Method for Source Localization in Bearing-Only Sensor Networks
In this study, the source localization problem is investigated using a group of networked sensors if only bearing measurements are available. The existing bearing-only source localization methods are subjected to a critical problem, i.e., how to achieve a fine balance between the computational complexity and the positioning accuracy when the threshold effect is considered. A novel range-selective generalized weighted centroid localization (RSGWCL) estimator is proposed to address the above-mentioned problem. Without the threshold effect, the proposed RSGWCL method is capable of achieving comparable positioning accuracy to the maximum likelihood estimator (MLE). If the threshold effect is identified, the RSGWCL estimator still has a capability of providing reliable estimates and outperforms the MLE. The result of the simulation is analyzed, such that the performance of the proposed method can be verified.
- 1 November 2021
- Published by Elsevier in Aerospace Science and Technology
- Vol. 118
Joint estimation of target location and relative altitude from angle measurements
- 1 November 2021
- Published by Elsevier in Aerospace Science and Technology
- Vol. 118
Joint estimation of target location and relative altitude from angle measurements
- 30 October 2021
- Published by Springer Nature
- p. 91-100
Bearing-Only Ground Slow Target Localization with Unknown Relative Altitude
- 30 October 2021
- Published by Springer Nature
- p. 91-100
Bearing-Only Ground Slow Target Localization with Unknown Relative Altitude