Two-dimensional endoscopic images are susceptible to interferences such as specular reflections and monotonous texture illumination,hindering accurate three-dimensional lesion reconstruction by surgical robots.This st...Two-dimensional endoscopic images are susceptible to interferences such as specular reflections and monotonous texture illumination,hindering accurate three-dimensional lesion reconstruction by surgical robots.This study proposes a novel end-to-end disparity estimation model to address these challenges.Our approach combines a Pseudo-Siamese neural network architecture with pyramid dilated convolutions,integrating multi-scale image information to enhance robustness against lighting interferences.This study introduces a Pseudo-Siamese structure-based disparity regression model that simplifies left-right image comparison,improving accuracy and efficiency.The model was evaluated using a dataset of stereo endoscopic videos captured by the Da Vinci surgical robot,comprising simulated silicone heart sequences and real heart video data.Experimental results demonstrate significant improvement in the network’s resistance to lighting interference without substantially increasing parameters.Moreover,the model exhibited faster convergence during training,contributing to overall performance enhancement.This study advances endoscopic image processing accuracy and has potential implications for surgical robot applications in complex environments.展开更多
Premise:The com bined effects of modern healthcare practices which prolong lifespan and declining birthrates have created unprecedented changes in age demographics worldwide that are especially pronounced in Japan,Sou...Premise:The com bined effects of modern healthcare practices which prolong lifespan and declining birthrates have created unprecedented changes in age demographics worldwide that are especially pronounced in Japan,South Korea,Europe,and North America.Since old age is the most significant predictor of dementia,global healthcare systems must rise to the challenge of providing care for those with neurodegenerative disorders.展开更多
Developing sensorless techniques for estimating battery expansion is essential for effective mechanical state monitoring,improving the accuracy of digital twin simulation and abnormality detection.Therefore,this paper...Developing sensorless techniques for estimating battery expansion is essential for effective mechanical state monitoring,improving the accuracy of digital twin simulation and abnormality detection.Therefore,this paper presents a data-driven approach to expansion estimation using electromechanical coupled models with machine learning.The proposed method integrates reduced-order impedance models with data-driven mechanical models,coupling the electrochemical and mechanical states through the state of charge(SOC)and mechanical pressure within a state estimation framework.The coupling relationship was established through experimental insights into pressure-related impedance parameters and the nonlinear mechanical behavior with SOC and pressure.The data-driven model was interpreted by introducing a novel swelling coefficient defined by component stiffnesses to capture the nonlinear mechanical behavior across various mechanical constraints.Sensitivity analysis of the impedance model shows that updating model parameters with pressure can reduce the mean absolute error of simulated voltage by 20 mV and SOC estimation error by 2%.The results demonstrate the model's estimation capabilities,achieving a root mean square error of less than 1 kPa when the maximum expansion force is from 30 kPa to 120 kPa,outperforming calibrated stiffness models and other machine learning techniques.The model's robustness and generalizability are further supported by its effective handling of SOC estimation and pressure measurement errors.This work highlights the importance of the proposed framework in enhancing state estimation and fault diagnosis for lithium-ion batteries.展开更多
The burgeoning market for lithium-ion batteries has stimulated a growing need for more reliable battery performance monitoring. Accurate state-of-health(SOH) estimation is critical for ensuring battery operational per...The burgeoning market for lithium-ion batteries has stimulated a growing need for more reliable battery performance monitoring. Accurate state-of-health(SOH) estimation is critical for ensuring battery operational performance. Despite numerous data-driven methods reported in existing research for battery SOH estimation, these methods often exhibit inconsistent performance across different application scenarios. To address this issue and overcome the performance limitations of individual data-driven models,integrating multiple models for SOH estimation has received considerable attention. Ensemble learning(EL) typically leverages the strengths of multiple base models to achieve more robust and accurate outputs. However, the lack of a clear review of current research hinders the further development of ensemble methods in SOH estimation. Therefore, this paper comprehensively reviews multi-model ensemble learning methods for battery SOH estimation. First, existing ensemble methods are systematically categorized into 6 classes based on their combination strategies. Different realizations and underlying connections are meticulously analyzed for each category of EL methods, highlighting distinctions, innovations, and typical applications. Subsequently, these ensemble methods are comprehensively compared in terms of base models, combination strategies, and publication trends. Evaluations across 6 dimensions underscore the outstanding performance of stacking-based ensemble methods. Following this, these ensemble methods are further inspected from the perspectives of weighted ensemble and diversity, aiming to inspire potential approaches for enhancing ensemble performance. Moreover, addressing challenges such as base model selection, measuring model robustness and uncertainty, and interpretability of ensemble models in practical applications is emphasized. Finally, future research prospects are outlined, specifically noting that deep learning ensemble is poised to advance ensemble methods for battery SOH estimation. The convergence of advanced machine learning with ensemble learning is anticipated to yield valuable avenues for research. Accelerated research in ensemble learning holds promising prospects for achieving more accurate and reliable battery SOH estimation under real-world conditions.展开更多
Interference significantly impacts the performance of the Global Navigation Satellite Systems(GNSS),highlighting the need for advanced interference localization technology to bolster anti-interference and defense capa...Interference significantly impacts the performance of the Global Navigation Satellite Systems(GNSS),highlighting the need for advanced interference localization technology to bolster anti-interference and defense capabilities.The Uniform Circular Array(UCA)enables concurrent estimation of the Direction of Arrival(DOA)in both azimuth and elevation.Given the paramount importance of stability and real-time performance in interference localization,this work proposes an innovative approach to reduce the complexity and increase the robustness of the DOA estimation.The proposed method reduces computational complexity by selecting a reduced number of array elements to reconstruct a non-uniform sparse array from a UCA.To ensure DOA estimation accuracy,minimizing the Cramér-Rao Bound(CRB)is the objective,and the Spatial Correlation Coefficient(SCC)is incorporated as a constraint to mitigate side-lobe.The optimization model is a quadratic fractional model,which is solved by Semi-Definite Relaxation(SDR).When the array has perturbations,the mathematical expressions for CRB and SCC are re-derived to enhance the robustness of the reconstructed array.Simulation and hardware experiments validate the effectiveness of the proposed method in estimating interference DOA,showing high robustness and reductions in hardware and computational costs associated with DOA estimation.展开更多
Previous multi-view 3D human pose estimation methods neither correlate different human joints in each view nor model learnable correlations between the same joints in different views explicitly,meaning that skeleton s...Previous multi-view 3D human pose estimation methods neither correlate different human joints in each view nor model learnable correlations between the same joints in different views explicitly,meaning that skeleton structure information is not utilized and multi-view pose information is not completely fused.Moreover,existing graph convolutional operations do not consider the specificity of different joints and different views of pose information when processing skeleton graphs,making the correlation weights between nodes in the graph and their neighborhood nodes shared.Existing Graph Convolutional Networks(GCNs)cannot extract global and deeplevel skeleton structure information and view correlations efficiently.To solve these problems,pre-estimated multiview 2D poses are designed as a multi-view skeleton graph to fuse skeleton priors and view correlations explicitly to process occlusion problem,with the skeleton-edge and symmetry-edge representing the structure correlations between adjacent joints in each viewof skeleton graph and the view-edge representing the view correlations between the same joints in different views.To make graph convolution operation mine elaborate and sufficient skeleton structure information and view correlations,different correlation weights are assigned to different categories of neighborhood nodes and further assigned to each node in the graph.Based on the graph convolution operation proposed above,a Residual Graph Convolution(RGC)module is designed as the basic module to be combined with the simplified Hourglass architecture to construct the Hourglass-GCN as our 3D pose estimation network.Hourglass-GCNwith a symmetrical and concise architecture processes three scales ofmulti-viewskeleton graphs to extract local-to-global scale and shallow-to-deep level skeleton features efficiently.Experimental results on common large 3D pose dataset Human3.6M and MPI-INF-3DHP show that Hourglass-GCN outperforms some excellent methods in 3D pose estimation accuracy.展开更多
The emergence of next generation networks(NextG),including 5G and beyond,is reshaping the technological landscape of cellular and mobile networks.These networks are sufficiently scaled to interconnect billions of user...The emergence of next generation networks(NextG),including 5G and beyond,is reshaping the technological landscape of cellular and mobile networks.These networks are sufficiently scaled to interconnect billions of users and devices.Researchers in academia and industry are focusing on technological advancements to achieve highspeed transmission,cell planning,and latency reduction to facilitate emerging applications such as virtual reality,the metaverse,smart cities,smart health,and autonomous vehicles.NextG continuously improves its network functionality to support these applications.Multiple input multiple output(MIMO)technology offers spectral efficiency,dependability,and overall performance in conjunctionwithNextG.This article proposes a secure channel estimation technique in MIMO topology using a norm-estimation model to provide comprehensive insights into protecting NextG network components against adversarial attacks.The technique aims to create long-lasting and secure NextG networks using this extended approach.The viability of MIMO applications and modern AI-driven methodologies to combat cybersecurity threats are explored in this research.Moreover,the proposed model demonstrates high performance in terms of reliability and accuracy,with a 20%reduction in the MalOut-RealOut-Diff metric compared to existing state-of-the-art techniques.展开更多
The reuse of liquid propellant rocket engines has increased the difficulty of their control and estimation.State and parameter Moving Horizon Estimation(MHE)is an optimization-based strategy that provides the necessar...The reuse of liquid propellant rocket engines has increased the difficulty of their control and estimation.State and parameter Moving Horizon Estimation(MHE)is an optimization-based strategy that provides the necessary information for model predictive control.Despite the many advantages of MHE,long computation time has limited its applications for system-level models of liquid propellant rocket engines.To address this issue,we propose an asynchronous MHE method called advanced-multi-step MHE with Noise Covariance Estimation(amsMHE-NCE).This method computes the MHE problem asynchronously to obtain the states and parameters and can be applied to multi-threaded computations.In the background,the state and covariance estimation optimization problems are computed using multiple sampling times.In real-time,sensitivity is used to quickly approximate state and parameter estimates.A covariance estimation method is developed using sensitivity to avoid redundant MHE problem calculations in case of sensor degradation during engine reuse.The amsMHE-NCE is validated through three cases based on the space shuttle main engine system-level model,and we demonstrate that it can provide more accurate real-time estimates of states and parameters compared to other commonly used estimation methods.展开更多
The beyond fifth-generation Internet of Things requires more capable channel coding schemes to achieve high-reliability,low-complexity and lowlatency communications.The theoretical analysis of error-correction perform...The beyond fifth-generation Internet of Things requires more capable channel coding schemes to achieve high-reliability,low-complexity and lowlatency communications.The theoretical analysis of error-correction performance of channel coding functions as a significant way of optimizing the transmission reliability and efficiency.In this paper,the efficient estimation methods of the block error rate(BLER)performance for rate-compatible polar codes(RCPC)are proposed under several scenarios.Firstly,the BLER performance of RCPC is generally evaluated in the additive white Gaussian noise channels.That is further extended into the Rayleigh fading channel case using an equivalent estimation method.Moreover,with respect to the powerful decoder such as successive cancellation list decoding,the performance estimation is derived analytically based on the polar weight spectrum and BLER upper bounds.Theoretical evaluation and numerical simulation results show that the estimated performance can fit well the practical simulated results of RCPC under the objective conditions,verifying the validity of our proposed performance estimation methods.Furthermore,the application designs of the reliability estimation of RCPC are explored,particularly in the advantages of the signal-to-noise(SNR)estimation and throughput efficiency optimization of polar coded hybrid automatic repeat request.展开更多
Precise and robust three-dimensional object detection(3DOD)presents a promising opportunity in the field of mobile robot(MR)navigation.Monocular 3DOD techniques typically involve extending existing twodimensional obje...Precise and robust three-dimensional object detection(3DOD)presents a promising opportunity in the field of mobile robot(MR)navigation.Monocular 3DOD techniques typically involve extending existing twodimensional object detection(2DOD)frameworks to predict the three-dimensional bounding box(3DBB)of objects captured in 2D RGB images.However,these methods often require multiple images,making them less feasible for various real-time scenarios.To address these challenges,the emergence of agile convolutional neural networks(CNNs)capable of inferring depth froma single image opens a new avenue for investigation.The paper proposes a novel ELDENet network designed to produce cost-effective 3DBounding Box Estimation(3D-BBE)froma single image.This novel framework comprises the PP-LCNet as the encoder and a fast convolutional decoder.Additionally,this integration includes a Squeeze-Exploit(SE)module utilizing the Math Kernel Library for Deep Neural Networks(MKLDNN)optimizer to enhance convolutional efficiency and streamline model size during effective training.Meanwhile,the proposed multi-scale sub-pixel decoder generates high-quality depth maps while maintaining a compact structure.Furthermore,the generated depthmaps provide a clear perspective with distance details of objects in the environment.These depth insights are combined with 2DOD for precise evaluation of 3D Bounding Boxes(3DBB),facilitating scene understanding and optimal route planning for mobile robots.Based on the estimated object center of the 3DBB,the Deep Reinforcement Learning(DRL)-based obstacle avoidance strategy for MRs is developed.Experimental results demonstrate that our model achieves state-of-the-art performance across three datasets:NYU-V2,KITTI,and Cityscapes.Overall,this framework shows significant potential for adaptation in intelligent mechatronic systems,particularly in developing knowledge-driven systems for mobile robot navigation.展开更多
基金Supported by Sichuan Science and Technology Program(2023YFSY0026,2023YFH0004)Supported by the Institute of Information&Communications Technology Planning&Evaluation(IITP)grant funded by the Korean government(MSIT)(No.RS-2022-00155885,Artificial Intelligence Convergence Innovation Human Resources Development(Hanyang University ERICA)).
文摘Two-dimensional endoscopic images are susceptible to interferences such as specular reflections and monotonous texture illumination,hindering accurate three-dimensional lesion reconstruction by surgical robots.This study proposes a novel end-to-end disparity estimation model to address these challenges.Our approach combines a Pseudo-Siamese neural network architecture with pyramid dilated convolutions,integrating multi-scale image information to enhance robustness against lighting interferences.This study introduces a Pseudo-Siamese structure-based disparity regression model that simplifies left-right image comparison,improving accuracy and efficiency.The model was evaluated using a dataset of stereo endoscopic videos captured by the Da Vinci surgical robot,comprising simulated silicone heart sequences and real heart video data.Experimental results demonstrate significant improvement in the network’s resistance to lighting interference without substantially increasing parameters.Moreover,the model exhibited faster convergence during training,contributing to overall performance enhancement.This study advances endoscopic image processing accuracy and has potential implications for surgical robot applications in complex environments.
基金funded by the Natural Sciences and Engineering Research Council of Canada(RGPIN:2016-05964&2023-04283 to JHK)the University of Manitoba Tri-Agency Bridge Funding(#57289 to JHK)the Ricard Foundation’s Baxter Bursary(to JP)。
文摘Premise:The com bined effects of modern healthcare practices which prolong lifespan and declining birthrates have created unprecedented changes in age demographics worldwide that are especially pronounced in Japan,South Korea,Europe,and North America.Since old age is the most significant predictor of dementia,global healthcare systems must rise to the challenge of providing care for those with neurodegenerative disorders.
基金Fund supported this work for Excellent Youth Scholars of China(Grant No.52222708)the National Natural Science Foundation of China(Grant No.51977007)+1 种基金Part of this work is supported by the research project“SPEED”(03XP0585)at RWTH Aachen Universityfunded by the German Federal Ministry of Education and Research(BMBF)。
文摘Developing sensorless techniques for estimating battery expansion is essential for effective mechanical state monitoring,improving the accuracy of digital twin simulation and abnormality detection.Therefore,this paper presents a data-driven approach to expansion estimation using electromechanical coupled models with machine learning.The proposed method integrates reduced-order impedance models with data-driven mechanical models,coupling the electrochemical and mechanical states through the state of charge(SOC)and mechanical pressure within a state estimation framework.The coupling relationship was established through experimental insights into pressure-related impedance parameters and the nonlinear mechanical behavior with SOC and pressure.The data-driven model was interpreted by introducing a novel swelling coefficient defined by component stiffnesses to capture the nonlinear mechanical behavior across various mechanical constraints.Sensitivity analysis of the impedance model shows that updating model parameters with pressure can reduce the mean absolute error of simulated voltage by 20 mV and SOC estimation error by 2%.The results demonstrate the model's estimation capabilities,achieving a root mean square error of less than 1 kPa when the maximum expansion force is from 30 kPa to 120 kPa,outperforming calibrated stiffness models and other machine learning techniques.The model's robustness and generalizability are further supported by its effective handling of SOC estimation and pressure measurement errors.This work highlights the importance of the proposed framework in enhancing state estimation and fault diagnosis for lithium-ion batteries.
基金National Natural Science Foundation of China (52075420)Fundamental Research Funds for the Central Universities (xzy022023049)National Key Research and Development Program of China (2023YFB3408600)。
文摘The burgeoning market for lithium-ion batteries has stimulated a growing need for more reliable battery performance monitoring. Accurate state-of-health(SOH) estimation is critical for ensuring battery operational performance. Despite numerous data-driven methods reported in existing research for battery SOH estimation, these methods often exhibit inconsistent performance across different application scenarios. To address this issue and overcome the performance limitations of individual data-driven models,integrating multiple models for SOH estimation has received considerable attention. Ensemble learning(EL) typically leverages the strengths of multiple base models to achieve more robust and accurate outputs. However, the lack of a clear review of current research hinders the further development of ensemble methods in SOH estimation. Therefore, this paper comprehensively reviews multi-model ensemble learning methods for battery SOH estimation. First, existing ensemble methods are systematically categorized into 6 classes based on their combination strategies. Different realizations and underlying connections are meticulously analyzed for each category of EL methods, highlighting distinctions, innovations, and typical applications. Subsequently, these ensemble methods are comprehensively compared in terms of base models, combination strategies, and publication trends. Evaluations across 6 dimensions underscore the outstanding performance of stacking-based ensemble methods. Following this, these ensemble methods are further inspected from the perspectives of weighted ensemble and diversity, aiming to inspire potential approaches for enhancing ensemble performance. Moreover, addressing challenges such as base model selection, measuring model robustness and uncertainty, and interpretability of ensemble models in practical applications is emphasized. Finally, future research prospects are outlined, specifically noting that deep learning ensemble is poised to advance ensemble methods for battery SOH estimation. The convergence of advanced machine learning with ensemble learning is anticipated to yield valuable avenues for research. Accelerated research in ensemble learning holds promising prospects for achieving more accurate and reliable battery SOH estimation under real-world conditions.
基金the financial support from the National Key Research and Development Program of China(No.2023YFB3907001)the National Natural Science Foundation of China(Nos.U2233217,62371029)the UK Engineering and Physical Sciences Research Council(EPSRC),China(Nos.EP/M026981/1,EP/T021063/1 and EP/T024917/)。
文摘Interference significantly impacts the performance of the Global Navigation Satellite Systems(GNSS),highlighting the need for advanced interference localization technology to bolster anti-interference and defense capabilities.The Uniform Circular Array(UCA)enables concurrent estimation of the Direction of Arrival(DOA)in both azimuth and elevation.Given the paramount importance of stability and real-time performance in interference localization,this work proposes an innovative approach to reduce the complexity and increase the robustness of the DOA estimation.The proposed method reduces computational complexity by selecting a reduced number of array elements to reconstruct a non-uniform sparse array from a UCA.To ensure DOA estimation accuracy,minimizing the Cramér-Rao Bound(CRB)is the objective,and the Spatial Correlation Coefficient(SCC)is incorporated as a constraint to mitigate side-lobe.The optimization model is a quadratic fractional model,which is solved by Semi-Definite Relaxation(SDR).When the array has perturbations,the mathematical expressions for CRB and SCC are re-derived to enhance the robustness of the reconstructed array.Simulation and hardware experiments validate the effectiveness of the proposed method in estimating interference DOA,showing high robustness and reductions in hardware and computational costs associated with DOA estimation.
基金supported in part by the National Natural Science Foundation of China under Grants 61973065,U20A20197,61973063.
文摘Previous multi-view 3D human pose estimation methods neither correlate different human joints in each view nor model learnable correlations between the same joints in different views explicitly,meaning that skeleton structure information is not utilized and multi-view pose information is not completely fused.Moreover,existing graph convolutional operations do not consider the specificity of different joints and different views of pose information when processing skeleton graphs,making the correlation weights between nodes in the graph and their neighborhood nodes shared.Existing Graph Convolutional Networks(GCNs)cannot extract global and deeplevel skeleton structure information and view correlations efficiently.To solve these problems,pre-estimated multiview 2D poses are designed as a multi-view skeleton graph to fuse skeleton priors and view correlations explicitly to process occlusion problem,with the skeleton-edge and symmetry-edge representing the structure correlations between adjacent joints in each viewof skeleton graph and the view-edge representing the view correlations between the same joints in different views.To make graph convolution operation mine elaborate and sufficient skeleton structure information and view correlations,different correlation weights are assigned to different categories of neighborhood nodes and further assigned to each node in the graph.Based on the graph convolution operation proposed above,a Residual Graph Convolution(RGC)module is designed as the basic module to be combined with the simplified Hourglass architecture to construct the Hourglass-GCN as our 3D pose estimation network.Hourglass-GCNwith a symmetrical and concise architecture processes three scales ofmulti-viewskeleton graphs to extract local-to-global scale and shallow-to-deep level skeleton features efficiently.Experimental results on common large 3D pose dataset Human3.6M and MPI-INF-3DHP show that Hourglass-GCN outperforms some excellent methods in 3D pose estimation accuracy.
基金funding from King Saud University through Researchers Supporting Project number(RSP2024R387),King Saud University,Riyadh,Saudi Arabia.
文摘The emergence of next generation networks(NextG),including 5G and beyond,is reshaping the technological landscape of cellular and mobile networks.These networks are sufficiently scaled to interconnect billions of users and devices.Researchers in academia and industry are focusing on technological advancements to achieve highspeed transmission,cell planning,and latency reduction to facilitate emerging applications such as virtual reality,the metaverse,smart cities,smart health,and autonomous vehicles.NextG continuously improves its network functionality to support these applications.Multiple input multiple output(MIMO)technology offers spectral efficiency,dependability,and overall performance in conjunctionwithNextG.This article proposes a secure channel estimation technique in MIMO topology using a norm-estimation model to provide comprehensive insights into protecting NextG network components against adversarial attacks.The technique aims to create long-lasting and secure NextG networks using this extended approach.The viability of MIMO applications and modern AI-driven methodologies to combat cybersecurity threats are explored in this research.Moreover,the proposed model demonstrates high performance in terms of reliability and accuracy,with a 20%reduction in the MalOut-RealOut-Diff metric compared to existing state-of-the-art techniques.
基金supported by the National Natural Science Foundation of China(Nos.62120106003 and 62173301)。
文摘The reuse of liquid propellant rocket engines has increased the difficulty of their control and estimation.State and parameter Moving Horizon Estimation(MHE)is an optimization-based strategy that provides the necessary information for model predictive control.Despite the many advantages of MHE,long computation time has limited its applications for system-level models of liquid propellant rocket engines.To address this issue,we propose an asynchronous MHE method called advanced-multi-step MHE with Noise Covariance Estimation(amsMHE-NCE).This method computes the MHE problem asynchronously to obtain the states and parameters and can be applied to multi-threaded computations.In the background,the state and covariance estimation optimization problems are computed using multiple sampling times.In real-time,sensitivity is used to quickly approximate state and parameter estimates.A covariance estimation method is developed using sensitivity to avoid redundant MHE problem calculations in case of sensor degradation during engine reuse.The amsMHE-NCE is validated through three cases based on the space shuttle main engine system-level model,and we demonstrate that it can provide more accurate real-time estimates of states and parameters compared to other commonly used estimation methods.
基金supported by National Natural Science Foundation of China(No.62201596)Research Planning Project of National University of Defense Technology(ZK22-45).
文摘The beyond fifth-generation Internet of Things requires more capable channel coding schemes to achieve high-reliability,low-complexity and lowlatency communications.The theoretical analysis of error-correction performance of channel coding functions as a significant way of optimizing the transmission reliability and efficiency.In this paper,the efficient estimation methods of the block error rate(BLER)performance for rate-compatible polar codes(RCPC)are proposed under several scenarios.Firstly,the BLER performance of RCPC is generally evaluated in the additive white Gaussian noise channels.That is further extended into the Rayleigh fading channel case using an equivalent estimation method.Moreover,with respect to the powerful decoder such as successive cancellation list decoding,the performance estimation is derived analytically based on the polar weight spectrum and BLER upper bounds.Theoretical evaluation and numerical simulation results show that the estimated performance can fit well the practical simulated results of RCPC under the objective conditions,verifying the validity of our proposed performance estimation methods.Furthermore,the application designs of the reliability estimation of RCPC are explored,particularly in the advantages of the signal-to-noise(SNR)estimation and throughput efficiency optimization of polar coded hybrid automatic repeat request.
文摘Precise and robust three-dimensional object detection(3DOD)presents a promising opportunity in the field of mobile robot(MR)navigation.Monocular 3DOD techniques typically involve extending existing twodimensional object detection(2DOD)frameworks to predict the three-dimensional bounding box(3DBB)of objects captured in 2D RGB images.However,these methods often require multiple images,making them less feasible for various real-time scenarios.To address these challenges,the emergence of agile convolutional neural networks(CNNs)capable of inferring depth froma single image opens a new avenue for investigation.The paper proposes a novel ELDENet network designed to produce cost-effective 3DBounding Box Estimation(3D-BBE)froma single image.This novel framework comprises the PP-LCNet as the encoder and a fast convolutional decoder.Additionally,this integration includes a Squeeze-Exploit(SE)module utilizing the Math Kernel Library for Deep Neural Networks(MKLDNN)optimizer to enhance convolutional efficiency and streamline model size during effective training.Meanwhile,the proposed multi-scale sub-pixel decoder generates high-quality depth maps while maintaining a compact structure.Furthermore,the generated depthmaps provide a clear perspective with distance details of objects in the environment.These depth insights are combined with 2DOD for precise evaluation of 3D Bounding Boxes(3DBB),facilitating scene understanding and optimal route planning for mobile robots.Based on the estimated object center of the 3DBB,the Deep Reinforcement Learning(DRL)-based obstacle avoidance strategy for MRs is developed.Experimental results demonstrate that our model achieves state-of-the-art performance across three datasets:NYU-V2,KITTI,and Cityscapes.Overall,this framework shows significant potential for adaptation in intelligent mechatronic systems,particularly in developing knowledge-driven systems for mobile robot navigation.
