Safe motion planning in robotics requires planning into space which has been verified to be free of obstacles. However, obtaining such environment representations using lidars is challenging by virtue of the sparsity of their depth measurements. We present a learning-aided 3D lidar reconstruction framework that upsamples sparse lidar depth measurements with the aid of overlapping camera images so as to generate denser reconstructions with more definitively free space than can be achieved with the raw lidar measurements alone. We use a neural network with an encoder-decoder structure to predict dense depth images along with depth uncertainty estimates which are fused using a volumetric mapping system. We conduct experiments on real-world outdoor datasets captured using a handheld sensing device and a legged robot. Using input data from a 16-beam lidar mapping a building network, our experiments showed that the amount of estimated free space was increased by more than 40% with our approach. We also show that our approach trained on a synthetic dataset generalises well to real-world outdoor scenes without additional fine-tuning. Finally, we demonstrate how motion planning tasks can benefit from these denser reconstructions.

We present an efficient, elastic 3D LiDAR reconstruction framework which can reconstruct up to maximum LiDAR ranges (60 m) at multiple frames per second, thus enabling robot exploration in large-scale environments. Our approach only requires a CPU. We focus on three main challenges of large-scale reconstruction: integration of long-range LiDAR scans at high frequency, the capacity to deform the reconstruction after loop closures are detected, and scalability for long-duration exploration. Our system extends upon a state-of-the-art efficient RGB-D volumetric reconstruction technique, called supereight, to support LiDAR scans and a newly developed submapping technique to allow for dynamic correction of the 3D reconstruction. We then introduce a novel pose graph clustering and submap fusion feature to make the proposed system more scalable for large environments. We evaluate the performance using two public datasets including outdoor exploration with a handheld device and a drone, and with a mobile robot exploring an underground room network. Experimental results demonstrate that our system can reconstruct at 3 Hz with 60 m sensor range and ∼5 cm resolution, while state-of-the-art approaches can only reconstruct to 25 cm resolution or 20 m range at the same frequency.

Terrorists often take the chemical clusters as the attacking target because of the adverse impacts of a chemical accident on society and the environment. In addition to some fixed countermeasures, previous studies have verified the feasibility of a patrol in addressing adversarial attacks. However, the previous patrolling practices fail to tackle the terrorist attacking problems in a large-scale area cost-effectively. To further tackle the protection issue with cost-beneficial solutions in a large-scale scenario, i.e., in a chemical cluster, we propose an area-partition-involved collaborative patrolling (APCP) game. We first leverage the proposed greedy deployment algorithm to determine the initial deployment of defenders (patrollers), including the quantity and position of patrol vehicles. Then, the large-scale area is partitioned into multiple smaller areas by using the collaborative idea of static partitioning. In the meantime, corresponding patrolling graphs are constructed based on graphic modeling methods. Finally, the APCP game is built between patrol vehicles (namely defender) and potential terrorists (namely attacker), in which patrol vehicles aim at detecting attack behaviors of terrorists by intelligently scheduling the patrolling routes. After formalizing the problem into a sequential game, we compute the Stackelberg equilibrium through the MultiLPs algorithm. Through case studies of three practical chemical cluster scenarios, the results explicitly show the superiority of our proposed APCP game by saving up to 25.48 % patrolling costs in a one-shot game compared to the results before partition. As for the collaborative patrolling problem in a large-scale area, the methods and models proposed in this paper can facilitate the management department of chemical clusters with intelligently scheduled patrolling routes, which can effectively reduce the cost of patrollers, and better protect the chemical cluster.

Estimating gas source terms is essential and significant for managing a gas emission accident. Optimization method, as a kind of estimation methods, is helpful to figure out the source terms by solving the inverse problem. Significantly, the performance of optimization method on source term estimation is affected by the accuracy of forward dispersion model. To enhance the estimation accuracy, previous works have demonstrated the feasibility of using Back Propagation Neural Network (BPNN) trained by actual experimental datasets as a forward dispersion model. However, the overall accuracy of source estimation is still limited by backward estimation methods. Most related studies used a single optimization algorithm to estimate source terms, which usually fails to realize the requirements of both high calculation accuracy and satisfying computational efficiency. Therefore, a hybrid strategy was proposed in this study to combine optimization algorithms with different characteristics, including particle swarm optimization, genetic algorithm and simulated annealing algorithm, to not only achieve high accuracy in global searching, but also converge to a stable result efficiently. Finally, extensive experiments are conducted to testify our proposed hybrid optimization algorithms. The Skill scores of hybrid optimization algorithms decrease obviously compared to those of single optimization algorithm. Hence, the proposed hybrid strategy is potentially useful for guiding the combination of optimization algorithms for gas source terms estimation, which further contributes to deal with a gas emission accident with satisfying calculation accuracy and computational efficiency.

Dispersion prediction plays a significant role in the management and emergency response to hazardous gas emissions and accidental leaks. Compared with conventional atmospheric dispersion models, machine leaning (ML) models have both high accuracy and efficiency in terms of prediction, especially in field cases. However, selection of model type and the inputs of the ML model are still essential problems. To address this issue, two ML models (i.e., the back propagation (BP) network and support vector regression (SVR) with different input selections (i.e., original monitoring parameters and integrated Gaussian parameters) are proposed in this paper. To compare the performances of presented ML models in field cases, these models are evaluated using the Prairie Grass and Indianapolis field data sets. The influence of the training set scale on the performances of ML models is analyzed as well. Results demonstrate that the integrated Gaussian parameters indeed improve the prediction accuracy in the Prairie Grass case. However, they do not make much difference in the Indianapolis case due to their inadaptability to the complex terrain conditions. In addition, it can be summarized that the SVR shows better generalization ability with relatively small training sets, but tends to under-fit the training data. In contrast, the BP network has a stronger fitting ability, but sometimes suffers from an over-fitting problem. As a result, the model and input selection presented in this paper will be of great help to environmental and public health protection in real applications.