Qiang Liu

Accurate LiDAR–camera extrinsic calibration is a prerequisite for multi-sensor fusion in autonomous driving. Existing online calibration methods often use monocular pseudo-depth to bridge the 2D–3D modality gap. However, the frequent degradation of pseudo-depth estimates in visually ambiguous regions inevitably induces scale ambiguity. Furthermore, single-stage architectures inherently struggle to reconcile the competing demands of wide-range pose search and sub-degree boundary alignment. This often leaves residual errors, especially in the pitch angle. To address these limitations, we propose UnCalib, an uncertainty-guided cascaded cross-modal calibration framework. We construct a heteroscedastic probabilistic depth prior to estimate pixel-wise depth reliability and reduce scale drift. Then, we use a matrix-based S E ( 3 ) pose update to prevent geometric distortion and orthogonality violations caused by linear pose updates in Euclidean space. In addition, we design a coarse-to-fine cascaded optimization strategy with edge-aware supervision to address weak gradients during fine pitch-angle alignment. Experiments show that UnCalib maintains stable convergence even under extreme initial perturbations of ± 1.5 m in translation and ± 20 ∘ in rotation. Our method achieves a mean rotation error of 0.033 ∘ and a mean translation error of 0.08 cm, providing a robust calibration solution for multi-sensor fusion.

Optimal operation control (OOC) for complex industrial processes (such as mineral flotation) is often constrained by multi-timescale dynamics and model-plant mismatch. From an artificial intelligence perspective, this paper proposes a novel data-driven hierarchical reinforcement learning (HRL) architecture. Through a state-dependent adaptive switching mechanism and a dual-scale integral compensation strategy, it effectively decouples the complex control task into two time-scale independent loops—solving the stability and plasticity dilemma in policy optimization. In engineering applications, the proposed framework is specifically tailored to the high-precision requirements of mineral flotation processes. At the upper layer, a proximal policy optimization (PPO) agent optimizes slow-scale economies of scale, while the lower layer employs a twin delayed deep deterministic policy gradient (TD3) controller, achieving robust and fast setpoint tracking under strict actuator constraints. Simulation results in a coarser flotation process validate the effectiveness of the proposed architecture, demonstrating significant improvements in tracking accuracy and operational robustness.

This article provides practical guidance for engineering managers and policymakers to design cost-efficient and sustainable REVB-RNs. Engineering managers should adopt multi-period network structure, incorporate low-carbon regulations and address heterogeneity of REVB capacities through a collaborative framework that integrates the EVM's proprietary recycling network with 3PRs. Specifically, embedding WD of CRA at the tactical level with REVB-RN design at the strategic level can minimize the total cost under uncertainty. Capacity-based categorization of REVBs enables more specialized and granular management of REVB-RNs. For policymakers, effective inducements for responsible recycling include setting appropriate carbon caps and taxes and imposing penalties for unrecycled REVBs. Overall, the proposed collaborative framework can support a more sustainable and cost-efficient ecosystem for related departments to better recycle REVBs. This study contributes to SDGs 9, 12 and 13.

Autonomous vehicles and robots work in a dynamic environment, which includes complex urban streets, dynamic obstacles, and complex sensing environments. It makes the perception task more challenging. A single type of sensor alone cannot meet the needs of target detection. Multimodal sensor fusion, which combines LiDAR and camera modalities, provides complementary 2D semantic and 3D geometric information. The performance of multimodal sensor fusion critically depends on precise extrinsic calibration between sensors. We propose CMTNet, a novel cross-modal Transformer architecture for robust extrinsic parameters estimation. The method uses depth maps as a unified representation of images and LiDAR point clouds. We utilize the ResNet-18 network to extract relative depth and semantic features from the monocular depth map. From the point cloud depth map, we extract precise 3D geometric features. Then, the correlation layer fuses the two features. Finally, the transformer estimates accurate calibration parameters based on multimodal features. We evaluated our method on the KITTI raw dataset, and it outperformed other methods. In addition, extensive experiments evaluating the model on KITTI odometry demonstrated that our method exhibited well generalization ability.

Multifactory remanufacturing is a widely adopted sustainable manufacturing approach nowadays. Its complexity lies in coordinating the dismantling, remanufacturing, and resource circulation among factories to maximize resource reuse and minimize environmental impact. Proper worker scheduling is crucial in this process to ensure efficient workflow and maximal resource utilization. This study proposes and addresses an optimization problem for multifactory remanufacturing considering worker scheduling, which is mainly divided into two parts: worker scheduling and remanufacturing process optimization (MRPO). A mixed-integer programming (MIP) mathematical model is established with the objective of profit maximization. A discrete battle royale optimization (BRO) algorithm is proposed to solve this problem, with a novel encoding structure and three soldier search strategies devised to better search for the optimal solution. The correctness of the model is validated through experiments on cases of different scales and comparisons with the IBM CPLEX optimizer. Furthermore, comparisons with the carnivorous plant algorithm (CPA), whale optimization algorithm (WAO), dingo optimization algorithm, and migrating bird optimization algorithm demonstrate the superiority and effectiveness of the proposed algorithm.

Robot path planning plays a critical role in enhancing the efficiency and accuracy of inspections and ultimately contributes to the safety of chemical production. In particular, the performance of robot inspection is influenced by amount of gas leakage and path length. In this study, a chemical safety inspection path multi-objective optimization model is constructed that considers minimizing path length and maximizing gas leakage detection sensitivity. To address this model, an improved multi-objective discrete growth optimization algorithm is proposed, which adopts discrete operators to meet the solution requirements of the discrete model and an adaptive leader selection strategy to further improve the algorithm solution performance. Finally, numerical computations and simulation experiments are conducted to validate the feasibility of the proposed method for inspection path planning. The results demonstrate that the method outperforms multi-objective discrete particle swarm optimization, non-dominated sorting genetic algorithm-II, and multi-objective gray wolf optimization in terms of solution convergence and distribution uniformity. Moreover, it can provide better non-dominated solutions in simulation experiments and provide multiple references for practical applications.