In this paper, we propose a deep reinforcement learning (DRL)-based method that allows unmanned aerial vehicles (UAVs) to execute navigation tasks in large-scale complex environments. This technique ...is important for many applications such as goods delivery and remote surveillance. The problem is formulated as a partially observable Markov decision process (POMDP) and solved by a novel online DRL algorithm designed based on two strictly proved policy gradient theorems within the actor-critic framework. In contrast to conventional simultaneous localization and mapping-based or sensing and avoidance-based approaches, our method directly maps UAVs' raw sensory measurements into control signals for navigation. Experiment results demonstrate that our method can enable UAVs to autonomously perform navigation in a virtual large-scale complex environment and can be generalized to more complex, larger-scale, and three-dimensional environments. Besides, the proposed online DRL algorithm addressing POMDPs outperforms the state-of-the-art.
•A model is proposed to describe the water-heat coupling process during freezing.•Unfrozen water function for saline soil is determined and applied in modeling.•Frozen depth of the saline region is ...predicted based on the coupling analysis.•Finale ice formation and water dynamics are simulated and validated.•Temperature variation hysteresis in soil profile is proved and analysed.
Frozen depth has a great significance for the foundation engineering in cold regions, always showing a high correlation with some attendant engineering phenomena, including water aggregation, frost heave, and salt accumulation. To study the heat-water dynamics and frozen depth characteristics during the freezing process, soils in western Jilin Province of China, a typical seasonal frozen region, were selected for investigation. A coupled heat and water model was proposed to describe the water-heat coupling process during freezing, with full consideration of the unfrozen water variation, the ice layer formation, and the interaction among different elements. Then, the dynamics of the heat-water and frozen depth were simulated based on the boundary conditions of temperature variation with reference to the meteorological data. The in- situ monitoring data from the whole winter were used to analyse the model performance. The results show that water content and temperature data match the test data, and the Root Mean Square Error (RMSE) values of the temperatures (within 2 °C) at different depths were acceptable, indicating that the water-heat dynamics can predict the maximum frozen depth well. In addition, the temperature of the soil profile varies rapidly in the first 60 days of winter, and the frozen depth continues to increase even though the temperature starts to rise after freezing for 80 days. The moisture transfers upwards with the effect of heat flow, and the formation of ice occurred mainly at a depth of 1.5 m. Heat conduction plays an important role in modelling, predominantly leading to the hysteresis in the frozen depth variation during freezing. This new method can provide a reference for water-heat movement and the prediction of the frozen depth during freezing in the saline soil regions.
The facultative anaerobe Salmonella strain VNP20009 selectively colonizes into tumors following systemic injection due to its preference for the hypoxia in the tumor cores. However, the phase 1 ...clinical trial of VNP20009 has been terminated mainly due to its weak antitumor effects and exhibition of dose-dependent toxicity. Here, we leveraged the advantages of VNP20009 biotherapy together with polydopamine-mediated photothermal therapy in order to enhance the antitumor efficacy toward malignant melanoma. VNP20009 was coated with polydopamine via oxidation and self-polymerization, which was then injected into tumor-bearing mice via the tail vein. Polydopamine-coated VNP20009 targeted hypoxic areas of the solid tumors, and near-infrared laser irradiation of the tumors induced heating due to polydopamine. This combined approach eliminated the tumors without relapse or metastasis with only one injection and laser irradiation. More importantly, we found both VNP and pDA potentiate the therapeutic ability of each other, resulting in a superior anticancer effect.
•A new method for detecting the position of cows’ eyes and udders is proposed.•The best position for the head to make eye temperature more accurate is recognized.•The enhanced infrared image more ...accurately recognizes the eyes and udders of cows.•This method improves the detection accuracy of key parts of dairy cows.
Mastitis is a disease that affects dairy cow health, and the timely detection of mastitis can improve the efficiency of mastitis treatment and reduce economic losses in the milk industry. To improve the detection speed and achieve automatic recognition of dairy cow mastitis, this study proposed a deep learning network EFMYOLOv3 (Enhanced Fusion MobileNetV3 You Only Look Once v3) based on the bilateral filtering enhancement of thermal images. EFMYOLOv3 is used to automatically detect dairy cow eyes and udders and is applied to the detection of mastitis in dairy cows based on thermal infrared images. We proposed a bilateral filtering image enhancement algorithm based on gray histograms to enhance image details to compensate for weak thermal image details and enhance the contrast between the foreground and background. We chose the lightweight MobileNetV3 as the backbone of YOLOv3. Based on the location attention mechanism, we used the multiscale enhanced fusion feature pyramid network structure as the feature extraction module. The feature map used for prediction was designed with the appropriate resolution and powerful multilayer semantic features to improve the accuracy of target detection. We replaced the standard convolutions in the base layer with depthwise separable convolutions to reduce the number of learning parameters. To verify the effectiveness of the target detection algorithm, the accuracy, recall, average frame rate, average accuracy and other indicators were compared with the SSD (single shot multibox detector) and YOLOv3 (You Only Look Once v3) algorithms. The test results revealed that the average frame rate of the EFMYOLOv3 algorithm is 99 frames per second (fps), and the average accuracy is 96.8%, which means that the key parts of the cow can be detected quickly and accurately. The temperature difference between the eyes and the udders was obtained by the target detection algorithm, and the mastitis detection of dairy cows was performed and compared with the somatic cell count (SCC). The results showed that the accuracy of the mastitis classification algorithm is 83.33%, and the sensitivity and specificity are 92.31% and 76.47%, respectively. This method realized accurate positioning of key parts of dairy cows and can be used for the automatic recognition of dairy cow mastitis.
Electric vehicles with a distributed drive train configuration provide great possibilities for the improvement of the vehicle dynamics, handling, safety as well as efficiency. In this paper, an ...energy-efficient torque allocation scheme is proposed for the improvement of traction efficiency and braking energy recovery. In traction conditions, the traction distribution is developed using an objective function of minimizing power loss of four electric motors. In braking conditions, aiming at guaranteeing the braking stability and recapturing the braking energy as much as possible, the changeable distribution of braking torque is obtained based on the ideal front-rear braking force distribution curve, while complying with braking regulations of Economic Commission for Europe. The proposed allocation scheme does not rely on the complex online computation. It is obtained via an offline optimization procedure and utilized for online allocation by simple interpolation. The low calculation effort makes it easy to implement the algorithm on real vehicles. Additionally, a conventional torque allocation is introduced as a contrasting approach. Finally, the simulations are conducted in CarSim and MATLAB/Simulink environment. The results demonstrate that the energy-efficient torque allocation scheme considerably improves the vehicle efficiency and increases the braking energy recovery compared with the conventional approach.
Smart manufacturing is the core idea of the fourth industrial evolution. For a smart manufacturing shop floor, real-time monitoring, simulation and prediction of manufacturing operations are vital to ...improve the production efficiency and flexibility. In this paper, the Cyber-Physical System (CPS) and Digital Twin technologies are introduced to build the interconnection and interoperability of a physical shop floor and corresponding cybershop floor. A Digital Twin-based Cyber-Physical Production System (DT-CPPS) is further established, and the configuring mechanism, operating mechanism and real-time data-driven operations control of DT-CPPS are discussed in detail. It is expected that DT-CPPS will provide the basis for shop floors to march towards smart manufacturing.
Increasing evidence indicates that metabolic disorders in offspring can result from the father's diet, but the mechanism remains unclear. In a paternal mouse model given a high-fat diet (HFD), we ...showed that a subset of sperm transfer RNA–derived small RNAs (tsRNAs), mainly from 5’ transfer RNA halves and ranging in size from 30 to 34 nucleotides, exhibited changes in expression profiles and RNA modifications. Injection of sperm tsRNA fractions from HFD males into normal zygotes generated metabolic disorders in the F₁ offspring and altered gene expression of metabolic pathways in early embryos and islets of F₁ offspring, which was unrelated to DNA methylation at CpG-enriched regions. Hence, sperm tsRNAs represent a paternal epigenetic factor that may mediate intergenerational inheritance of diet-induced metabolic disorders.
2D black phosphorus (BP) nanomaterials are presented as a delivery platform. The endocytosis pathways and biological activities of PEGylated BP nanosheets in cancer cells are revealed for the first ...time. Finally, a triple‐response combined therapy strategy is achieved by PEGylated BP nanosheets, showing a promising and enhanced antitumor effect.
To achieve on‐demand drug release, mesoporous silica nanocarriers as antitumor platforms generally need to be gated with stimuli‐responsive capping agents. Herein, a “smart” mesoporous nanocarrier ...that is gated by the drug itself through a pH‐sensitive dynamic benzoic–imine covalent bond is demonstrated. The new system, which tactfully bypasses the use of auxiliary capping agents, could also exhibit desirable drug release at tumor tissues/cells and enhanced tumor inhibition. Moreover, a facile dynamic PEGylation via benzoic–imine bond further endows the drug‐self‐gated nanocarrier with tumor extracellular pH‐triggered cell uptake and improves therapeutic efficiency in vivo. In short, the paradigm shift in capping agents here will simplify mesoporous nanomaterials as intelligent drug carriers for cancer therapy. Moreover, the self‐gated strategy in this work also shows general potential for self‐controlled delivery of natural biomolecules, for example, DNA/RNA, peptides, and proteins, due to their intrinsic amino groups.
A drug‐self‐gated strategy for mesoporous nanocarrier could achieve on‐demand drug release at tumor tissue/cells and improved antitumor efficiency. The key is using a pH‐sensitive benzoic–imine bond for dynamic conjugation of amino‐containing drug molecules (i.e., doxorubicin) on the pore outlets.
Lead halide perovskites have been promising materials for lasing applications. Despite that a series of perovskite microlasers have been reported, their lasing modes are confined by either the ...as-grown morphology or the etched boundary. The first one is quite random and incompatible with integration, whereas the latter one strongly spoils the laser performances. Herein, we propose and experimentally demonstrate a robust and generic mechanism to realize well-controlled perovskite microlasers without the etching process. By patterning a one-dimensional polymer grating onto a perovskite film, we show that the symmetry-protected bound states in the continuum (BICs) can be formed in it. The intriguing properties of BICs including a widely spread mode profile and high Q factor, associated with the exceptional gain of perovskite, produce single-mode microlasers with high repeatability, controllability, directionality, and a polarization vortex. This mechanism can also be extended to two-dimensional nanostructures, enabling BIC lasers with different topological charges.