In the field of satellite imagery, remote sensing image captioning (RSIC) is a hot topic with challenge overfitting and difficulty text alignment. To address these issues, this paper proposes vision-language aligning paradigm for RSIC to jointly represent vision language. First, new dataset DIOR-Captions built augmenting object detection in optical (DIOR) images manually annotated Chinese English contents. Second, Vision-Language model Cross-modal Attention (VLCA) presented generate accurate...

Unmanned ground vehicles (UGVs) and unmanned aerial (UAVs) have been widely used in delivery. In the context of COVID-2019, order to control development epidemic, many places adopted measures isolate close area once a confirmed case is found. While reducing contact between people, it also blocks normal driving vehicles. Only by changing traditional logistics distribution methods can customers who closed isolated for long time be served. Therefore, we use Cooperative UGV-UAV achieve it. this...

Hexagonal grids use a hierarchical subdivision tessellation to cover the entire plane or sphere. Due 6-fold rotational symmetry, hexagonal have some advantages (e.g. isoperimetry, equidistant neighbors, and uniform connectivity) over quadrangular triangular girds, which makes them suitable tackle tasks of geospatial information processing intelligent decision-making. In this paper, we first introduce applications based on grids. Then, planer spherical analyze group representations for them,...

Vision-and-Language Navigation (VLN) is a task designed to enable embodied agents carry out natural language instructions in realistic environments. Most VLN tasks, however, are guided by an elaborate set of that depicted step-by-step. This approach deviates from real-world problems which humans only describe the object and its surroundings allow robot ask for help when required. Vision-based with Language-based Assistance (VNLA) recently proposed requires agent navigate find target...

Atom gravimeter (AG) is a kind of high-precision absolute based on atom interference, which can measure gravitational acceleration. Vibration one the important factors limits measurement accuracy. compensation an effective method to reduce impact vibration accuracy and it especially suitable for complex environments. In this article, improved proposed. Based simplified model transfer function, searches optimal parameters through chaotic sparrow search algorithm (CSSA) effectively restore...

Ship detection for high-resolution optical remote sensing (HRORS) imagery is widely applicable to Navy construction, maritime transportation, safety, port management, and many other fields. Optical remote-sensing images can easily be disturbed by thin cloud, light, some complex environmental conditions, which makes the task of ship challenging. In this paper, we propose a new general network framework recognize ships with arbitrary-orientation, in use novel way define rotating rectangular...

In the field of robotics, it is crucial to obtain a comprehensive semantic understanding scene for many applications. Based on behavioral topological map and graph, we propose employ named Topological Scene Map (TSM) representation in indoor environment understanding. The constructed expresses spatial connection relations semantically describes navigation behavior between adjacent nodes. graph promotes TSM record objects that appear objects. addition makes expression more specific, which...

This study addressed a problem of rapid velocity consensus within swarm unmanned aerial vehicles. Our analytical framework was based on tools using matrix theory and algebraic graph theory. We established connections between connectivity the speed converging velocity. The relationship communication cost established. To deal with trade-off among connectivity, convergence cost, we propose distributed small world network construction method. characteristics expedite toward in vehicle swarm....

Satellite communication systems are increasingly facing serious environmental challenges such as malicious jamming, monitoring, and intercepting. As a current development of artificial intelligence, intelligent jammers with learning ability can effectively perceive the surrounding spectrum environment to dynamically change their jamming strategies. result, mainstream satellite anti-jamming technology based on wide interval high-speed frequency hopping is unable deal this problem effectively....

Autonomous navigation is essential to promoting the autonomy of aircraft. While adjusting velocity direction and altitude angles navigate goal, aircraft must maintain speed prevent losing height. In this research, a deep reinforcement learning method with potential-based dense reward function proposed teach agent independently. The integrates help PPO algorithm task. Experiments are conducted among three different functions: sparse reward, reward. results show that can rarely guide navigate,...

In cognitive radio systems, there are cases where malicious users transmit interference power to prevent secondary from transmitting information. Due the destructive behavior of users, spectrum utilization efficiency will be reduced. The applications game theory study this confrontation relationship reasonable. Based on Continuous Blotto Game (CBG) model under condition one-shot perfect information confrontation, paper constructs an anti-jamming for allocation between and simulates...

The interactive instruction following task requires intelligent robots to complete natural language instructions by interacting with the surrounding environment through visual perception and mechanical manipulation. most representative benchmark is ALFRED challenge task, which aims perform restricted sequential actions accomplish daily housework in a virtual environment. However, recently proposed multi-modal planning approaches that combine vision do not well on task. In this work, we...

Goal recognition (GR) is a method of inferring the goals other agents, which enables humans or AI agents to proactively make response plans. design (GRD) has been proposed deliberately redesign underlying environment accelerate goal recognition. Along with GR and GRD problems, in this paper, we start by introducing control (GRC) problem under network interdiction, focuses on controlling process. When observer attempts facilitate explainability actor’s behavior reducing uncertainty, actor...

Unmanned ground vehicles (UGVs) have been widely used in security patrol. The existence of two potential opponents, the malicious teammate (cooperative) and hostile observer (adversarial), highlights importance privacy-preserving planning under contested environments. In a cooperative setting, disclosure private information can be restricted to teammates. adversarial obfuscation added control observability observer. this paper, we attempt generate opponent-aware plans, mainly focusing on...

The problem of adversarial Unmanned ground vehicles (UGVs) patrol has received increasing attention in recent decade. UGV maneuver may encounter one fullknowledge opponent the setting, which makes it difficult to ensure safety. In this paper, we propose framework based on generative networks (GAN) generate reliable paths and recommend top-k deception score by employing some deceptive tactics during path evaluation process. our proposed provides with an that helps from local position desired...

With the development of science and technology, environment where Unmanned Ground Vehicles (UGVs) patrol is complex for cooperative adversarial interactions. The actions performed by UGVs can be observed others which might cooperators or adversaries. In order to fulfill battle field task, goals path planning should controlled in mixed environment. this paper, we first define goal control problem among three different roles (actor, observer, observer), actor (UGV) need reveal real tied with...

Inspired by the dream of using natural language to guide an Agent, we propose a understanding model that links commands with particular target states, so agent can better understand tasks expressed in language. The performs exploration environment according human intentions through decision-making method reinforcement learning, which will improve performance and efficiency exploration. By most advanced text-based pre-trained Bert, is robust commands.

Knowing the enemy and yourself, you can fight a hundred battles win them all. For an unknown opponent, nash equilibrium may be useful strategy to not lose on worst-case situation. However, when we have opponent's estimated model, developing exploitive counter-strategy wins far beyond equilibrium. But how take performance maximizing keep least best response value, is critical compute robust counter strategy. In this paper, try achieve higher expected payoff by modeling stationary opponent...

Recent successes in human-computer gaming AI development have achieved super-human performance sophisticated confrontation games. Since the Nash equilibrium of multiplayer games is difficult to solve even within only two players, poker has always been an challenging problem field Artificial Intelligence. From perspective game theory, pursuing diverse polices necessary non-stationary environment, which deeply rooted non-transitive structure game. In this paper, order provide a tractable...

Camouflaged object is usually visually similar to the surrounding environment. Although camouflaged detection more challenging than general detection, has been used in medicine, biology and military, so research on also of great significance.Since camouflage can be regarded as "salient object" image, this paper combines characteristics salient mainly from three aspects edge information enhancement, channel attention mechanism expansion receptive field. Inspired by idea U-Net, we design a new...

Urban combat environment is affected by buildings and traffic network with certain spatial characteristics. A single unmanned vehicle carries out the coverage reconnaissance task will be constrained limited information low efficiency, so multiple UGVs(Unmanned Ground Vehicles) cooperative becomes an effective means to solve problem. In this paper, urban formulated as a routing problem, new traversal model proposed, which takes account connectivity characteristics of road. At same time,...

Network structures with strong heterogeneity will be more helpful to the spread of cooperative strategies in evolutionary process. After increasing benefits betrayal strategies, networks better resist intrusion betrayers and maintain a higher level cooperation. In this paper, we introduce degree factor into individual payoff add heterogeneity, so that individuals high degrees obtain benefits, while income low reduced. way, between enhanced during evolution network, network can betrayers.