Prior Knowledge Of Environment Research Articles

GPLM: Enhancing underwater images with Global Pyramid Linear Modulation

Underwater imagery often suffers from challenges such as color distortion, low contrast, blurring, and noise due to the absorption and scattering of light in water. These degradations complicate visual interpretation and hinder subsequent image processing. Existing methods struggle to effectively address the complex, spatially varying degradations without prior environmental knowledge or may produce unnatural enhancements.To overcome these limitations, we propose a novel method called Global Pyramid Linear Modulation that integrates physical degradation modeling with deep learning for underwater image enhancement. Our approach extends Feature-wise Linear Modulation to a four-dimensional structure, enabling fine-grained, spatially adaptive modulation of feature maps. Our method captures multi-scale contextual information by incorporating a feature pyramid architecture with self-attention and feature fusion mechanisms, effectively modeling complex degradations. We validate our method by integrating it into the MixDehazeNet model and conducting experiments on benchmark datasets. Our approach significantly improves the Peak Signal-to-Noise Ratio, increasing from 28.6 dB to 30.6 dB on the EUVP-515-test dataset. Compared to recent state-of-the-art methods, our method consistently outperforms them by over 3 dB in PSNR on datasets with ground truth. It improves the Underwater Image Quality Measure by more than one on datasets without ground truth. Furthermore, we demonstrate the practical applicability of our method on a real-world underwater dataset, achieving substantial improvements in image quality metrics and visually compelling results. These experiments confirm that our method effectively addresses the limitations of existing techniques by adaptively modeling complex underwater degradations, highlighting its potential for underwater image enhancement tasks.

Range-Domain Subspace Detector in the Presence of Direct Blast for Forward Scattering Detection in Shallow-Water Environments

This paper aims to detect a target that crosses the baseline connecting the source and the receiver in shallow-water environments, which is a special scenario for a bistatic sonar system. In such a detection scenario, an intense sound wave, known as the direct blast, propagates directly from the source to the receiver without target scattering. This direct blast usually arrives at the receiver simultaneously with the forward scattering signal and exhibits a larger intensity than the signal, posing a significant challenge for target detection. In this paper, a range-domain subspace is constructed by the horizontal distance between the source/target and each element of a horizontal linear array (HLA) when the ranges of environmental parameters are known a priori. Meanwhile, a range-domain subspace detector based on direct blast suppression (RSD-DS) is proposed for forward scattering detection. The source and the target are located at different positions, so the direct blast and the scattered signal are in different range-domain subspaces. By projecting the received data onto the orthogonal complement subspace of the direct blast subspace, the direct blast can be suppressed and the signal that lies outside the direct blast subspace is used for target detection. The simulation results indicate that the proposed RSD-DS exhibits a performance close to the generalized likelihood ratio detector (GLRD) while requiring less prior knowledge of environments (only known are the ranges of the sediment sound speed and the bottom sound speed), and its robustness to environmental uncertainties is better than that of the latter. Moreover, the proposed RSD-DS exhibits better immunity against the direct blast than the GLRD, since it can still work effectively at a signal-to-direct blast ratio (SDR) of −30 dB, while the GLRD stops working in this case.

Spatial domain decomposition-based physics-informed neural networks for practical acoustic propagation estimation under ocean dynamics.

Practical acoustic propagation modeling is significantly affected by ocean dynamics, and then can be exploited in numerous oceanic applications, where "practical" refers to modeling acoustic propagation in simulations that mimic real-world ocean environments. Physics-based numerical models provide approximate solutions of wave equation and rely on accurate prior environmental knowledge while the environment of practical scenarios is largely unknown. In contrast, data-driven machine learning offers a promising avenue to estimate practical acoustic propagation by learning from dataset. However, collecting such a high-quality, noise-free, and dense dataset remains challenging. Under the practical hurdle posed by the above approaches, the emergence of physics-informed neural network (PINN) presents an alternative to integrate physics and data but with limited representation capacity. In this work, a framework, termed spatial domain decomposition-based physics-informed neural networks (SPINNs), is proposed to enhance the representation capacity in spatially dependent oceanic scenarios and effectively learn from incomplete and biased prior physics and noisy dataset. Experiments demonstrate SPINNs' advantages over PINN in practical acoustic propagation estimation. The learning capacity of SPINNs toward physics and dataset during training is further analyzed. This work holds promise for practical applications and future expansion.

HRL-Based Access Control for Wireless Communications With Energy Harvesting

This paper studies the access control problem of long-term throughput maximization in wireless communication systems with Energy Harvesting (EH). In the existing research, many access schemes based on accurate environmental information have been proposed, such as channel information and the EH process. However, access to environmental information is costly, and traditional access control frameworks are expensive to explore in high-dimensional spaces. Thus, an access control framework based on hierarchical reinforcement learning (HRL) is proposed in this paper. In HRL, the control problem in the Markov decision process (MDP) form is decomposed into a multilevel sequential control problem. It includes high-level channel number selection, mid-level channel selection, and low-level channel matching subproblems. The scheme is obtained by combining the solutions of subproblems at different level which are solved in sequence. In addition, to improve learning efficiency, the deterministic action (DA) module and the prior knowledge (PK) module are put forward. The DA module solves the channel matching problem under the additional guidance given by the previous subproblem, which selects definite good low-level actions. The PK module provides the framework with the common knowledge of the system structure learned from the hypothetical environment, so as to obtain better initial performance. Experimental results show that our framework achieves better performance and better learning efficiency compared with several recent transmission schemes. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Note to Practitioners</i> —Access control is an important issue in wireless communication systems, and users need to be scheduled to solve the constraint of limited resources, such as energy usually provided by batteries. In recent years, in order to overcome the energy limitation, energy harvesting devices have been developed and applied to wireless communication systems. However, the energy collection ability of the system is greatly influenced by the environment, which leads to the poor performance of most traditional control schemes that rely on the prior knowledge of the environment. Therefore, this paper proposes a novel hierarchical reinforcement learning (HRL)-based model-free access control framework for wireless communication system to maximize the system throughput without any prior environmental knowledge. The scheme abstracts the original control problem into three sub-control sub control problems according to tasks and solves them sequentially, thus simplifying the original control problem. This scheme can not only learn independently, but also does not depend on the prior knowledge of the environment. Moreover, this method is also suitable for the large-scale environment while the conventional end-to-end reinforcement learning is not suitable for. Compared with traditional algorithms, our method has better performance and higher learning efficiency.

Particle Filtering for Source Depth and Water Depth Joint Tracking in Shallow Water

Environmental mismatch degrades the performance of source localization and tracking methods in shallow water. One solution is to estimate source parameters and the key environmental parameters simultaneously from the acoustic data. In this paper, an unconventional approach of joint tracking source depth and water depth parameters by a particle filter is proposed. This approach is free of prior environmental knowledge and numerical calculation of any forward model. First, a state-space model based on modal nature behavior is established driving the shallow-water propagation, instead of modeling in time or space, as was done previous works. Subsequently, particle filtering is employed for joint tracking, in which the evolution with mode-order of vertical wavenumbers and the relationship between state parameters and beam-wavenumber outputs transformed from the data are exploited. Final, the particle smoother reduces the uncertainty of state parameters at initial steps, and improves the overall tracking accuracy. Our approach is demonstrated using simulated data in an ideal waveguide and applied to shallow-water SWellEx-96 experimental data to substantiate its superior performance.

APD: Learning Diverse Behaviors for Reinforcement Learning Through Unsupervised Active Pre-Training

Unsupervised pre-training in reinforcement learning enables the agent to gain prior environmental knowledge, which is then fine-tuned in the supervised stage to quickly adapt to various downstream tasks. In the absence of task-related rewards, pre-training aims to acquire policies (i.e., behaviors) that generate different trajectories to explore and master the environment. Previous research categorizes states into their associated behaviors by learning a supervised discriminator. However, an underlying problem persists: such discriminator is trained in lack of relevant data, leading to an underestimation of reward for new states and inadequate exploration. To this end, we introduce an unsupervised active pre-training algorithm for diverse behavior induction (APD). We explicitly characterize the behavior variables with a state-dependent sampling method, and the agent can decompose the entire state space into parts for fine-grained and diverse behavior learning. Specifically, a particle-based entropy estimator is applied to optimize a combination of behavioral entropy and mutual information objective. Moreover, we develop behavior-based representation learning to compress states into the latent space. Experiments show that our method can improve exploration efficiency and outperforms most state-of-the-art unsupervised algorithms on a number of continuous control tasks in the DeepMind Control Suite.

Emergence of probabilistic representation in the neural network of primary visual cortex

SummaryDuring the early development of the mammalian visual system, the distribution of neuronal preferred orientations in the primary visual cortex (V1) gradually shifts to match major orientation features of the environment, achieving its optimal representation. By combining computational modeling and electrophysiological recording, we provide a circuit plasticity mechanism that underlies the developmental emergence of such matched representation in the visual cortical network. Specifically, in a canonical circuit of densely-interconnected pyramidal cells and inhibitory parvalbumin-expressing (PV+) fast-spiking interneurons in V1 layer 2/3, our model successfully simulates the experimental observations and further reveals that the nonuniform inhibition plays a key role in shaping the network representation through spike timing-dependent plasticity. The experimental results suggest that PV + interneurons in V1 are capable of providing nonuniform inhibition shortly after vision onset. Our study elucidates a circuit mechanism for acquisition of prior knowledge of environment for optimal inference in sensory neural systems

Use of multipath time-delay ratio for source depth estimation with a vertical line array in deep water.

In this paper, a method for the problem of depth estimation of a broadband source via reliable acoustic path propagation is presented for the case using a vertical line array (VLA). The estimates are determined by two kinds of multipath time-delay ratios, namely, the ratio of direct-surface-reflected (D-SR) to direct-direct time-delays and the ratio of D-SR to surface-reflected-surface-reflected time-delays. The innovation of ratio behavior is that it provides a mechanism for obtaining a useful depth interval with the assumption of plane-wave propagation. The estimation accuracy of a depth interval relies on the degree to which the actual acoustic propagation characteristic can be modeled by image theory. Furthermore, the variability of depth interval due to the approximation made in the derivation method allows one to achieve binary discrimination of both the source depth and source range with only a minimal amount of prior environmental knowledge. The methodology of multipath time-delay estimation is first reviewed and improved, followed by an illustration of the source depth estimation and a discussion of the performance analysis using results from numerical simulations. Finally, the proposed method is demonstrated with experimental data collected in the South China Sea in which a short-aperture VLA is deployed near the sea bottom.

Time Reversal Linearly Constrained Minimum Power Algorithm for Direction of Arrival Estimation in Diffuse Multipath Environments

Direction of arrival (DOA) estimation in diffuse multipath environments is a challenge for ground-based radar remote sensing applications, which has significant value in military fields, such as air defense surveillance. However, radar received echo usually contains various multipath signals caused by the reflection of complex ground or sea surface. With the introduction of multipath signals, traditional algorithms’ performance on angle estimation decreases severely. In response to this problem, the letter proposes a new time reversal (TR) algorithm used for multiple-input multiple-output (MIMO) radar angle estimation. First, the algorithm reconstructs a TR covariance matrix by multiplexing the data’s rows and columns, increasing the estimation accuracy of the TR covariance matrix. Besides, the letter applies a linearly constrained minimum power (LCMP) constraint to suppress diffuse multipath signals according to the prior knowledge of environments. Simulation results examine the improvement of estimation accuracy by the proposed algorithm, also verify the superiority of the proposed algorithm in different multipath scenarios. What’s more, the algorithm has broader applicability due to avoiding the difficulties of removing the coherence and estimating multipath number in practice.

Safeguard Network Slicing in 5G: A Learning Augmented Optimization Approach

Network slicing, as a key 5G enabling technology, is promising to support with more flexibility, agility, and intelligence towards the provisioned services and infrastructure management. Fulfilling these tasks is challenging, as nowadays networks are increasingly heterogeneous, dynamic and large-dimensioned. This contradicts the dominant network slicing solutions that only customize immediate performance over one snapshot of the system in the literature. Instead, this paper first presents a two-stage slicing optimization model with time-averaged metrics to safeguard the network slicing in the dynamical networks, where prior environmental knowledge is absent but can be partially observed at runtime. Directly solving an off-line solution to this problem is intractable since the future system realizations are unknown before decisions. Therefore, we propose a learning augmented optimization approach with deep learning and Lyapunov stability theories. This enables the system to learn a safe slicing solution from both historical records and run-time observations. We prove that the proposed solution is always feasible and nearly optimal, up to a constant additive factor. Finally, we demonstrate up to 2.6× improvement in the simulation when compared with three state-of-the-art algorithms.

Adaptive Beamforming with Inadequate Snapshots

In array signal processing, the covariance matrix used to calculate the adaptive weights is often poor estimated when the snapshot number is inadequate. The prior environmental knowledge can be used to make the estimation more accuracy. In this paper, an alternative knowledge-aided adaptive beamforming approach that is robust to low sample support environment is proposed. In this algorithm the covariance matrix used to calculate the optimum weights is constructed by blending a sample covariance matrix and a priori structured covariance matrix. Numerical simulations demonstrate the proposed algorithm has the potential for substantial performance improvement.

Bowhead whale localization using asynchronous hydrophones in the Chukchi Sea.

This paper estimates bowhead whale locations and uncertainties using non-linear Bayesian inversion of their modally-dispersed calls recorded on asynchronous recorders in the Chukchi Sea, Alaska. Bowhead calls were recorded on a cluster of 7 asynchronous ocean-bottom hydrophones that were separated by 0.5-9.2 km. A warping time-frequency analysis is used to extract relative mode arrival times as a function of frequency for nine frequency-modulated whale calls that dispersed in the shallow water environment. Each call was recorded on multiple hydrophones and the mode arrival times are inverted for: the whale location in the horizontal plane, source instantaneous frequency (IF), water sound-speed profile, seabed geoacoustic parameters, relative recorder clock drifts, and residual error standard deviations, all with estimated uncertainties. A simulation study shows that accurate prior environmental knowledge is not required for accurate localization as long as the inversion treats the environment as unknown. Joint inversion of multiple recorded calls is shown to substantially reduce uncertainties in location, source IF, and relative clock drift. Whale location uncertainties are estimated to be 30-160 m and relative clock drift uncertainties are 3-26 ms.

Environmental knowledge management: A long-term enabler of tourism development

A review of the literature confirms that there is a need for knowledge management frameworks which support knowledge creation, particularly in those organisations having to operate in a changing environment. Socialisation, externalisation, combination and internalisation continue to be key processes for the creation of knowledge which enables organisations to successfully address environmental challenges. This paper examines the relevance and importance of a SECI model as an enabler of the processes of reusing and updating the environmental knowledge of an organisation. The research reported has confirmed that time is a key component of the implementation of a SECI model in organisations operating in a changing environment. It has been confirmed that such a strategy should focus on the reuse of prior environmental knowledge as a mechanism to establish within the organisation a context where new knowledge management processes are understood and adopted by employees. The paper also examines the relationship between environmental knowledge and organisational performance indicators. These relationships are examined through an empirical study of 87 companies in the Spanish hospitality sector. The results of the study indicate that environmental knowledge at any given time (T) is significant in predicting the knowledge management processes that may be successfully implemented at a later point in time (T + 6 years).

Learning gait of quadruped robot without prior knowledge of the environment

Walking is the basic skill of a legged robot, and one of the promising ways to improve the walking performance and its adaptation to environment changes is to let the robot learn its walking by itself. Currently, most of the walking learning methods are based on robot vision system or some external sensing equipment to estimate the walking performance of certain walking parameters, and therefore are usually only applicable under laboratory condition, where environment can be pre-defined. Inspired by the rhythmic swing movement during walking of legged animals and the behavior of their adjusting their walking gait on different walking surfaces, a concept of walking rhythmic pattern(WRP) is proposed to evaluate the walking specialty of legged robot, which is just based on the walking dynamics of the robot. Based on the onboard acceleration sensor data, a method to calculate WRP using power spectrum in frequency domain and diverse smooth filters is also presented. Since the evaluation of WRP is only based on the walking dynamics data of the robot's body, the proposed method doesn't require prior knowledge of environment and thus can be applied in unknown environment. A gait learning approach of legged robots based on WRP and evolution algorithm(EA) is introduced. By using the proposed approach, a quadruped robot can learn its locomotion by its onboard sensing in an unknown environment, where the robot has no prior knowledge about this place. The experimental result proves proportional relationship exits between WRP match score and walking performance of legged robot, which can be used to evaluate the walking performance in walking optimization under unknown environment.

Improving navigation messages for mobile urban guides: Effects of the guide's interlocutor model, spatial abilities and use of landmarks on route description

The aim of this study was to define guidelines for designing dialogue-based pedestrian guidance applications. It was based on an investigation of human–human interactions over the phone, where one speaker took on the role of the guide and the other that of the person being guided. During these interactions, a concurrent navigation in a real large-scale environment was performed by the guided member of the dyad. The content of the route descriptions produced by the guides was analyzed in terms of the latter's representations of their interlocutors' prior knowledge of the route environment, their spatial abilities and the landmarks they mentioned. The main results show that when the guided subjects had no prior knowledge of the environment, guides with high spatial abilities provided more NLoc landmarks (e.g. “there is a bar”) than ExoLoc (e.g. “there is a bar to the right of the bank”) and EgoLoc landmarks (e.g. “there is a bar on your left”). Conversely, guides with low spatial abilities gave just as many NLoc landmarks as EgoLoc landmarks. Practical applications stemming from these results are discussed. Relevance to industry This paper assesses the usefulness of taking spatial abilities, prior knowledge and landmarks into account when designing a new type of guidance application, namely a dialogue-based pedestrian guidance system. It also indicates the different ways in which these factors can be taken into account during the development of such applications.

Knowledge-aided adaptive beamforming

In array processing, when the available snapshot number is comparable with or even smaller than the sensor number, the sample covariance matrix R̂ is a poor estimate of the true covariance matrix R. To estimate R more accurately, prior environmental knowledge can be used, which is manifested as knowing an a priori covariance matrix R0. In practice, R0 usually represents prior knowledge on dominant sources or interferences. Since the noise power level is unknown, and thus cannot be included into the a priori covariance matrix, R0 is often rank deficient. Both modified general linear combinations (MGLC) and modified convex combinations (MCC) of the a priori covariance matrix R0, the sample covariance matrix R̂ and an identity matrix I to obtain an enhanced estimate of R, denoted as R̃ are considered. Both MGLC and MCC can choose the combination weights fully automatically. Moreover, both the MGLC and MCC methods can be extended to deal with linear combinations of an arbitrary number of positive semi-definite matrices. Both approaches can be formulated as convex optimisation problems that can be solved efficiently to obtain globally optimal solutions. Numerical examples are provided to demonstrate the type of achievable performance by using R̃ instead of R̂ in the standard Capon beamformer.

Exploring importance of location and prior knowledge of environment on mobile robot control

Advances in technologies to create ever more sophisticated robots is outpacing our understanding of how such robots and humans successfully interact to accomplish specific tasks. In particular, it is now possible for humans to control the navigation of certain classes of robots via the Internet. Yet, we know little about how individual differences variables such as spatial ability and prior knowledge of the navigational space contribute to the success of human–robot interactions. Nor has there been much research on the kinds of interfaces that facilitate effective and efficient interaction, and whether these are differentially effective across the spectrum of spatial ability. In this study, an Internet control mechanism for an RWI B-14 mobile robot called “TourMate” was used to investigate the differences and similarities in how humans teleoperated the robot from two locations (intermediate and remote). Thirty-two participants who were assessed on spatial ability and prior knowledge of the space in which the robot was being controlled directed the robot in simple spatial navigation tasks. Time to complete the task, number of steps, and errors were collected. Results indicated that spatial ability and location, but not prior knowledge, were reliable predictor of efficiency of robot teleoperation. Information was also obtained about the utility of various features of the interface and suggested that individual difference characteristics of robot operators are related to the utility of different interface features. Implications for design of human–robot interaction systems are discussed.