Search

Multi-robot cooperative navigation is an important task, which has been widely studied in many fields like logistics, transportation, and disaster rescue. However, most of the existing methods either require some strong assumptions or are validated in simple scenarios, which greatly hinders their implementation in the real world. In this paper, more complex environments are considered in which robots can only acquire local observations from their own sensors and have only limited communication capabilities for mapless collaborative navigation. To address this challenging task, we propose a hierarchical framework, by fusing both Sensor-wise and Agent-wise features for Perception-Improving (SAPI), which can adaptively integrate features from different information sources to improve perception capabilities. Specifically, to facilitate scene understanding, we assign prior knowledge to the visual coder to generate efficient embeddings. For effective feature representation, an attention-based sensor fusion network is designed to fuse sensor-level information of visual and LiDAR sensors, while graph convolution with multi-head attention mechanism is applied to aggregate agent-level information from an arbitrary number of neighbors. In addition, reinforcement learning is used to optimize the policy, where a novel compound reward function is introduced to guide training. Extensive experiments demonstrate that our method has excellent generalization ability in different scenarios and scalability for large-scale systems.

Abrupt permafrost thaw accelerates the decomposition of soil organic carbon and might double the warming caused by the carbon release. However, the influence of thaw slump evolution on carbon dioxide (CO2) emission rates and its drives remains unclear, which induces large uncertainties in the prediction of permafrost carbon‐climate feedback. Here we collected soil samples in the thaw slump landscapes that happened 1–23 years ago on the northern Qinghai‐Tibet Plateau (QTP) and measured the CO2release rates using a 189‐day aerobic laboratory incubation in the dark. The incubation results showed that thaw slump occurred 23 years ago reduced soil CO2–C release by 57 ± 19% compared with the undisturbed area. The relative contribution of O‐alkyl C and microbial abundance decreases with the thaw slump initiation time lengthens. We illustrate that soil carbon quality and microbial communities uniquely explained 41% and 13% of the variation in CO2–C release, respectively. We preliminary estimate that the carbon release for thaw slump landscapes on the QTP may be overestimated by approximately 50% if the declining soil CO2–C release is without consideration. Our study highlights the CO2–C release would decrease with the stability of thaw slumps on the warming and wetting QTP, which may weaken the mountain permafrost carbon‐climate feedback. Rapid warming of permafrost regions leads to the widespread development of thaw slumps and accelerates the decomposition and release of soil organic carbon. Yet, how the evolution of thaw slump affects CO2release remains unclear. We found that CO2release gradually decreased with the thaw sequence. Labile carbon (O‐alkyl C) proportion and microbial abundance declined along the thaw sequence, while resistant carbon (alkyl C) proportion increased. Soil carbon quality and microbial communities are important factors controlling CO2release. Applying the results to the entire Tibetan Plateau, we find that permafrost carbon emissions would be overestimated by approximately 50% if the declining soil CO2release of stabilized thaw slump is without consideration. Our findings suggest that stable thaw slump can weaken CO2release. Therefore, the assessment of mountain permafrost carbon feedback needs to consider the effect of thaw slump evolution. Soil CO2release was reduced by 57% in areas that collapsed for 23 years compared to the control areasThaw slump leads to a large decrease in O‐alkyl C and soil microbial abundanceSoil carbon quality and microbial communities exert the controls of 41% and 13% over CO2release, respectively Soil CO2release was reduced by 57% in areas that collapsed for 23 years compared to the control areas Thaw slump leads to a large decrease in O‐alkyl C and soil microbial abundance Soil carbon quality and microbial communities exert the controls of 41% and 13% over CO2release, respectively

Tumor microparticles (T-MPs) are considered as a tumor vaccine candidate. Although some studies have analyzed the mechanism of T-MPs as tumor vaccine, we still lack understanding of how T-MPs stimulate a strong anti-tumor immune response. Here, we show that T-MPs induce macrophages to release a key chemotactic factor CCL2, which attracts monocytes to the vaccine injection site and enhances endocytosis of antigen. Monocytes subsequently enter the draining lymph node, and differentiate into monocyte-derived DCs (moDCs), which present tumor antigens to T lymphocytes and deliver a potent anti-tumor immune response. Mechanically, T-MPs activate the cGAS-STING signaling through DNA fragments, and then induce monocytes to upregulate the expression of IRF4, which is a key factor for monocyte differentiation into moDCs. More importantly, monocytes that have endocytosed T-MPs acquire the ability to treat tumors. Collectively, this work might provide novel vaccination strategy for the development of tumor vaccines and facilitate the application of T-MPs for clinic oncotherapy.

Laser cladding coating is the focus of development in the field of die repair. It has the characteristics of uniform and compact coating, good bonding strength, little thermal impact on the substrate, and easy to realize automation. FeCr alloy coating was prepared on 738H die steel by laser cladding process. Based on MSC Marc finite element software and life-and-death element method, the numerical analysis model of laser cladding repair was establelished, and the temperature field, stress field and deformation under different process parameters were solved. According to the simulation results, the cladding process was formulated, the hardness and linear thermal expansion coefficient of the cladding coating were characterized, and the effects of the process parameters on the structure and properties of the coating were studied. The cladding coating matched well with the substrate performance was obtained, and the high quality repair of the die surface was realized.

Laser cladding surface modification technology is widely used to strengthen and repair the surface of axially symmetrical parts. Based on the Gauss heat source model and COMSOL Multiphysics finite element analysis software, the surface modification process characteristics of laser cladding on the rotating substrate were studied by numerical simulation. The transient temperature field distribution under different process parameters was solved. Laser cladding pool corresponding to numerical simulation parameters was photographed by infrared high-speed camera. The results show that the temperature distribution of the Gauss heat source is delayed due to the normal distribution of heat flow and the cumulative effect of heat. In addition, with the increase of the rotational speed of the substrate, a cyclic preheating temperature band will be formed on the surface. The shape of molten pool obtained by temperature field analysis is droplet shape, which is basically consistent with the actual morphology of laser cladding molten pool.