Chenghu Cao
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
Research Subjects
Chenghu Cao has a total of 27 co-authors. Here are the top co-authors by number of shared publications:
- Yongbo Zhao
Xidian University
13 shared publications - Haisheng Huang
Xian Innovation College, Yanan University, Xian, China
3 shared publications - Junhua Yang
School of Electronic Engineering, Xi’an University of Posts and Telecommunications, Xi’an 710121, China
2 shared publications Keya Fu
Beihang University
2 shared publications- Santuan Qin
School of Electronic Engineering, Xi’an University of Posts and Telecommunications, Xi’an 710121, China
2 shared publications - Wei Cheng
Northwestern Polytechnical University
2 shared publications
Latest Publications
- Open Access
- 1 January 2026
- Published by Institution of Engineering and Technology (IET) in IET Radar, Sonar & Navigation
- Vol. 20 (1) , e70188
A Multi‐Scan Possibility GLMB Model for Multi‐Target Tracking Under Epistemic Uncertainty
This work addresses the difficult issue of tracking multiple targets in the presence of epistemic uncertainty. We propose a multiscan possibility‐based generalised labelled multiBernoulli (GLMB) smoothing approach designed to prevent track fragmentation and switching when detection profiles are unknown and vary over time. By incorporating a labelled uncertainty finite set (UFS) derived from possibility theory into the multiscan GLMB framework, the estimation of multitarget states is enhanced, yielding a possibility‐theoretic counterpart to the standard multiscan GLMB smoothing procedure. These multiscan possibility functions can effectively represent epistemic uncertainties that stem from limited knowledge of the detection profile. The prediction and update steps of these possibility functions are expressed using a Gaussian max‐mixture (GMM), which simplifies the execution of the multiscan possibility GLMB recursion. Simulation outcomes indicate that, under epistemic uncertainty, the proposed multiscan possibility GLMB smoothing provides improved estimates of both multitarget trajectories and target cardinality, particularly in practical scenarios.
- 1 January 2026
- Published by Institution of Engineering and Technology (IET) in IET Radar, Sonar & Navigation
- Vol. 20 (1) , e70188
Open AccessA Multi‐Scan Possibility GLMB Model for Multi‐Target Tracking Under Epistemic Uncertainty
This work addresses the difficult issue of tracking multiple targets in the presence of epistemic uncertainty. We propose a multiscan possibility‐based generalised labelled multiBernoulli (GLMB) smoothing approach designed to prevent track fragmentation and switching when detection profiles are unknown and vary over time. By incorporating a labelled uncertainty finite set (UFS) derived from possibility theory into the multiscan GLMB framework, the estimation of multitarget states is enhanced, yielding a possibility‐theoretic counterpart to the standard multiscan GLMB smoothing procedure. These multiscan possibility functions can effectively represent epistemic uncertainties that stem from limited knowledge of the detection profile. The prediction and update steps of these possibility functions are expressed using a Gaussian max‐mixture (GMM), which simplifies the execution of the multiscan possibility GLMB recursion. Simulation outcomes indicate that, under epistemic uncertainty, the proposed multiscan possibility GLMB smoothing provides improved estimates of both multitarget trajectories and target cardinality, particularly in practical scenarios.
- 7 November 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Vehicular Technology
- Vol. 75 (5) , 7557-7572
Joint DOA and DOD Estimation for Transmit Energy Focusing-Based Bistatic MIMO Radar Using Unitary HOSVD Method
In this paper, we consider the joint DOD and DOA estimation for transmit energy focusing-based (TEF) bistatic MIMO radar. We devise the optimization problem by minimizing the matching error between the designed transmit beampattern and the desired beampattern and using a re-weighted group sparsity norm to control the row sparsity of the transmit beamspace matrix. To solve the resultant nonconvex problem, the alternating direction method of multipliers (ADMM) framework is used to transform this nonconvex problem into multiple convex subproblems by introducing an auxiliary variables. To solve the non-differentiable of the re-weighted group sparse norm of the transmit beamspace matrix, we use the idea of proximal gradient (PG) to derive an iteration method to obtain the optimal solution of this subproblem. To exploit the structure inherent in multi-dimensional measurement data, we establish the tensor-based HOSVD model to jointly estimate DOD and DOA. The backward-forward averaging and spatial smoothing schemes are extended to tensor model mapping the centro-Hermitian matrices to real-valued ones with the same size. The unitary HOSVD algorithm in TEF-based bistatic MIMO radar is proposed to jointly estimate DOD and DOA of targets. Moreover, the proposed algorithm is suitable for correlated and closely-spaced targets. The simulation results validate the effectiveness of the proposed algorithm with respect to joint DOD and DOA estimation.
- 7 November 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Vehicular Technology
- Vol. 75 (5) , 7557-7572
Joint DOA and DOD Estimation for Transmit Energy Focusing-Based Bistatic MIMO Radar Using Unitary HOSVD Method
In this paper, we consider the joint DOD and DOA estimation for transmit energy focusing-based (TEF) bistatic MIMO radar. We devise the optimization problem by minimizing the matching error between the designed transmit beampattern and the desired beampattern and using a re-weighted group sparsity norm to control the row sparsity of the transmit beamspace matrix. To solve the resultant nonconvex problem, the alternating direction method of multipliers (ADMM) framework is used to transform this nonconvex problem into multiple convex subproblems by introducing an auxiliary variables. To solve the non-differentiable of the re-weighted group sparse norm of the transmit beamspace matrix, we use the idea of proximal gradient (PG) to derive an iteration method to obtain the optimal solution of this subproblem. To exploit the structure inherent in multi-dimensional measurement data, we establish the tensor-based HOSVD model to jointly estimate DOD and DOA. The backward-forward averaging and spatial smoothing schemes are extended to tensor model mapping the centro-Hermitian matrices to real-valued ones with the same size. The unitary HOSVD algorithm in TEF-based bistatic MIMO radar is proposed to jointly estimate DOD and DOA of targets. Moreover, the proposed algorithm is suitable for correlated and closely-spaced targets. The simulation results validate the effectiveness of the proposed algorithm with respect to joint DOD and DOA estimation.
- 1 September 2025
- Published by IOP Publishing in Engineering Research Express
- Vol. 7 (3) , 035289
A CSI fingerprint map-based indoor positioning algorithm using stacking fusion model for enhanced localization accuracy
The traditional indoor location sensing based on wireless signals is deeply affected by two factors, which are the impurity of source data and the singularity of location sensing model. To tackle this problem, we propose a Channel State Information (CSI) fingerprint sensing position algorithm based on Stacking Fusion Model (SFM) in this paper. Taking advantage of different data prediction and training models, we build a location sensing prediction model that combines Stacking ensemble learning with the corresponding machine learning algorithm and it is structured into two layers of learners. The SFM method is initially the first layer of base learners, four indoor location sensing algorithms: Support vector Regression (SVR), K-nearest Neighbor Regression (KNR), Random Forest (RF) and Weighted Mixed Regression based on SVR&KNR (WMR-SKR) are trained, and the prediction results are generated and output. Subsequently, the prediction results from the first layer are used as inputs by the second-layer learner, which trains a linear regression model. After the same processing, the CSI data of the sensing target is matched with the data output from the second layer of processing in the offline stage, thus completing the final location sensing result. We implement verification sensing experiments in two real indoor environments, and the results show that compared with only using the above four algorithms, the position sensing accuracy of SFM is improved 35.56%, 40.49%, 22.01%, 17.18% respectively in the laboratory environment, and by 38.61%, 59.34%, 34.74%, 23.46% respectively in the empty classroom environment. Considering the universality of experimental verification, in addition to the comparison with the four primary learners, the comparison with the external classical algorithm WKNN is introduced. The experimental results show that the positioning accuracy has also been further improved in the two environments. It should be mentioned that the above results can be achieved not only attributed to SFM, but also to the CSI data preprocessing algorithm proposed in this paper.
- 1 September 2025
- Published by IOP Publishing in Engineering Research Express
- Vol. 7 (3) , 035289
A CSI fingerprint map-based indoor positioning algorithm using stacking fusion model for enhanced localization accuracy
The traditional indoor location sensing based on wireless signals is deeply affected by two factors, which are the impurity of source data and the singularity of location sensing model. To tackle this problem, we propose a Channel State Information (CSI) fingerprint sensing position algorithm based on Stacking Fusion Model (SFM) in this paper. Taking advantage of different data prediction and training models, we build a location sensing prediction model that combines Stacking ensemble learning with the corresponding machine learning algorithm and it is structured into two layers of learners. The SFM method is initially the first layer of base learners, four indoor location sensing algorithms: Support vector Regression (SVR), K-nearest Neighbor Regression (KNR), Random Forest (RF) and Weighted Mixed Regression based on SVR&KNR (WMR-SKR) are trained, and the prediction results are generated and output. Subsequently, the prediction results from the first layer are used as inputs by the second-layer learner, which trains a linear regression model. After the same processing, the CSI data of the sensing target is matched with the data output from the second layer of processing in the offline stage, thus completing the final location sensing result. We implement verification sensing experiments in two real indoor environments, and the results show that compared with only using the above four algorithms, the position sensing accuracy of SFM is improved 35.56%, 40.49%, 22.01%, 17.18% respectively in the laboratory environment, and by 38.61%, 59.34%, 34.74%, 23.46% respectively in the empty classroom environment. Considering the universality of experimental verification, in addition to the comparison with the four primary learners, the comparison with the external classical algorithm WKNN is introduced. The experimental results show that the positioning accuracy has also been further improved in the two environments. It should be mentioned that the above results can be achieved not only attributed to SFM, but also to the CSI data preprocessing algorithm proposed in this paper.
- Open Access
Research on 3D Localization of Indoor UAV Based on Wasserstein GAN and Pseudo Fingerprint Map
In addition to outdoor environments, unmanned aerial vehicles (UAVs) also have a wide range of applications in indoor environments. The complex and changeable indoor environment and relatively small space make indoor localization of UAVs more difficult and urgent. An innovative 3D localization method for indoor UAVs using a Wasserstein generative adversarial network (WGAN) and a pseudo fingerprint map (PFM) is proposed in this paper. The primary aim is to enhance the localization accuracy and robustness in complex indoor environments. The proposed method integrates four classic matching localization algorithms with WGAN and PFM, demonstrating significant improvements in localization precision. Simulation results show that both the WGAN and PFM algorithms significantly reduce localization errors and enhance environmental adaptability and robustness in both small and large simulated indoor environments. The findings confirm the robustness and efficiency of the proposed method in real-world indoor localization scenarios. In the inertial measurement unit (IMU)-based tracking algorithm, using the fingerprint database of initial coarse particles and the fingerprint database processed by the WGAN algorithm to locate the UAV, the localization error of the four algorithms is reduced by 30.3% on average. After using the PFM algorithm for matching localization, the localization error of the UAV is reduced by 28% on average.
Open AccessResearch on 3D Localization of Indoor UAV Based on Wasserstein GAN and Pseudo Fingerprint Map
In addition to outdoor environments, unmanned aerial vehicles (UAVs) also have a wide range of applications in indoor environments. The complex and changeable indoor environment and relatively small space make indoor localization of UAVs more difficult and urgent. An innovative 3D localization method for indoor UAVs using a Wasserstein generative adversarial network (WGAN) and a pseudo fingerprint map (PFM) is proposed in this paper. The primary aim is to enhance the localization accuracy and robustness in complex indoor environments. The proposed method integrates four classic matching localization algorithms with WGAN and PFM, demonstrating significant improvements in localization precision. Simulation results show that both the WGAN and PFM algorithms significantly reduce localization errors and enhance environmental adaptability and robustness in both small and large simulated indoor environments. The findings confirm the robustness and efficiency of the proposed method in real-world indoor localization scenarios. In the inertial measurement unit (IMU)-based tracking algorithm, using the fingerprint database of initial coarse particles and the fingerprint database processed by the WGAN algorithm to locate the UAV, the localization error of the four algorithms is reduced by 30.3% on average. After using the PFM algorithm for matching localization, the localization error of the UAV is reduced by 28% on average.
- Open Access
- 30 November 2024
- Published by Institution of Engineering and Technology (IET) in IET Radar, Sonar & Navigation
- Vol. 18 (12) , 2698-2708
A belief propagation algorithm based on track‐before‐detect for tracking low‐observable and manoeuvering targets using multiple sensors
It is notoriously challenging work to track an unknown number of low‐observable manoeuvering targets. In this paper, a sequential Bayesian inference method based on the multiple‐model dynamic model and track‐before‐detect measurement (TBD) model is proposed for tracking low‐observable manoeuvering targets using multiple sensors. The multiple‐model dynamic model is capable to characterise the dynamic behaviour of manoeuvering targets. The TBD measurement model can completely capture an echo signal without any preprocessing, furtherly handling with low‐observable targets. The authors’ proposed method is based on a new multi‐sensor statistical model that allows targets to interact and contribute to more than one data cell for the pixeled image TBD approach. Based on the factor graph representing the multi‐sensor statistical model, the marginal posterior densities are derived by performing the message passing equations of the proposed belief propagation algorithm for target detection and target state estimation. The simulation results validate that the computational complexity of our proposed multi‐sensor BP‐TBD algorithm scales in the number of sensor nodes and demonstrate that its performance is superior among the state‐of‐the‐art multi‐sensor TBD methods.
- 30 November 2024
- Published by Institution of Engineering and Technology (IET) in IET Radar, Sonar & Navigation
- Vol. 18 (12) , 2698-2708
Open AccessA belief propagation algorithm based on track‐before‐detect for tracking low‐observable and manoeuvering targets using multiple sensors
It is notoriously challenging work to track an unknown number of low‐observable manoeuvering targets. In this paper, a sequential Bayesian inference method based on the multiple‐model dynamic model and track‐before‐detect measurement (TBD) model is proposed for tracking low‐observable manoeuvering targets using multiple sensors. The multiple‐model dynamic model is capable to characterise the dynamic behaviour of manoeuvering targets. The TBD measurement model can completely capture an echo signal without any preprocessing, furtherly handling with low‐observable targets. The authors’ proposed method is based on a new multi‐sensor statistical model that allows targets to interact and contribute to more than one data cell for the pixeled image TBD approach. Based on the factor graph representing the multi‐sensor statistical model, the marginal posterior densities are derived by performing the message passing equations of the proposed belief propagation algorithm for target detection and target state estimation. The simulation results validate that the computational complexity of our proposed multi‐sensor BP‐TBD algorithm scales in the number of sensor nodes and demonstrate that its performance is superior among the state‐of‐the‐art multi‐sensor TBD methods.
- 20 November 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Vehicular Technology
- Vol. 74 (3) , 4048-4063
Super-Resolution Range-Velocity Estimate of Multiple Targets for OFDM-Based 5G Radar Based on Unitary Parallel Factor Method
In this paper, we consider the problem of joint rangevelocity estimate of multiple targets in orthogonal frequency division multiplex (OFDM) transmit waveform-based 5G radar. A unitary parallel factor (PARAFAC) algorithm is proposed to achieve superresolution estimate and outstanding performance, using forwardbackward averaging scheme. The forward-backward averaging scheme is adopted to construct real-valued tensor signal model instead of the complex-valued one, yielding the better accuracy at modest complexity. The proposed unitary PARAFAC algorithm is performed by decomposing the real-valued tensor without signal subspace estimate. Due to the inherent smoothing processing of the proposed unitary PARAFAC algorithm, it can effectively deal with high correlated target. Additionally, the proposed unitary PARAFAC algorithm can automatically obtain pair parameters including range and velocity of the same target without additional pair-matching operation. More importantly the regularized alternative least squares (RALS) algorithm is used to improve the decomposition performance of the real-valued tensor while maintaining iteration stability. The numerical results are presented to demonstrate the superior performance, especially for high correlated and closely spaced targets in low-SNR scenario.
- 20 November 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Vehicular Technology
- Vol. 74 (3) , 4048-4063
Super-Resolution Range-Velocity Estimate of Multiple Targets for OFDM-Based 5G Radar Based on Unitary Parallel Factor Method
In this paper, we consider the problem of joint rangevelocity estimate of multiple targets in orthogonal frequency division multiplex (OFDM) transmit waveform-based 5G radar. A unitary parallel factor (PARAFAC) algorithm is proposed to achieve superresolution estimate and outstanding performance, using forwardbackward averaging scheme. The forward-backward averaging scheme is adopted to construct real-valued tensor signal model instead of the complex-valued one, yielding the better accuracy at modest complexity. The proposed unitary PARAFAC algorithm is performed by decomposing the real-valued tensor without signal subspace estimate. Due to the inherent smoothing processing of the proposed unitary PARAFAC algorithm, it can effectively deal with high correlated target. Additionally, the proposed unitary PARAFAC algorithm can automatically obtain pair parameters including range and velocity of the same target without additional pair-matching operation. More importantly the regularized alternative least squares (RALS) algorithm is used to improve the decomposition performance of the real-valued tensor while maintaining iteration stability. The numerical results are presented to demonstrate the superior performance, especially for high correlated and closely spaced targets in low-SNR scenario.
- 1 October 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in Journal of Systems Engineering and Electronics
- Vol. 35 (5) , 1109-1121
Multiple-Model GLMB Filter Based on Track-Before-Detect for Tracking Multiple Maneuvering Targets
- 1 October 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in Journal of Systems Engineering and Electronics
- Vol. 35 (5) , 1109-1121
Multiple-Model GLMB Filter Based on Track-Before-Detect for Tracking Multiple Maneuvering Targets