Incomplete Environmental Information Research Articles

ET-HF: A novel information sharing model to improve multi-agent cooperation

Many real-world multi-agent systems require agents to cooperate with each other. However, it is challenging to generate optimal cooperative strategies (e.g., location coordination or speed coordination in the encirclement task) in partially observable environments. Information sharing and temporal experience are two main ways to alleviate partial observability. The main idea of information sharing is observation-sharing scheme which is under-utilization of the topological information, hence agents obtain incomplete environmental information, resulting in a weak information sharing ability between agents. Our idea is to combine observation information and topology information to form a more comprehensive information sharing scheme. Inspired by this, we propose a novel information sharing model in multi-agent reinforcement learning (MARL), named the Entity-Teammate Hierarchical Fusion (ET-HF) model. It comprises three modules (1) message generation, using graph neural networks (GNNs) to combine local observations and topology information to generate two types of messages; (2) message interaction, transmitting messages, and adopting a hierarchical attention fusion mechanism to integrate the above two types of messages and finish information sharing; (3) cooperative policy optimization, inputting integrated messages into the proximal policy optimization (PPO) algorithm to generate the cooperative strategy that adapts to partial observability. Through ET-HF, agents can improve their understanding of the environment and reduce the loss of information caused by partial observability to generate a superior collaborative strategy. We train the multi-agent teams in a fully decentralized framework and the empirical results show that our model outperforms the baselines in cooperative tasks such as coverage control, formation control and line control.

Path Planning for Wheeled Mobile Robot in Partially Known Uneven Terrain.

Path planning for wheeled mobile robots on partially known uneven terrain is an open challenge since robot motions can be strongly influenced by terrain with incomplete environmental information such as locally detected obstacles and impassable terrain areas. This paper proposes a hierarchical path planning approach for a wheeled robot to move in a partially known uneven terrain. We first model the partially known uneven terrain environment respecting the terrain features, including the slope, step, and unevenness. Second, facilitated by the terrain model, we use A algorithm to plan a global path for the robot based on the partially known map. Finally, the Q-learning method is employed for local path planning to avoid locally detected obstacles in close range as well as impassable terrain areas when the robot tracks the global path. The simulation and experimental results show that the designed path planning approach provides satisfying paths that avoid locally detected obstacles and impassable areas in a partially known uneven terrain compared with the classical A algorithm and the artificial potential field method.

Changes of movement patterns from early dispersal to settlement

Moving and spatial learning are two intertwined processes: (a) changes in movement behavior determine the learning of the spatial environment, and (b) information plays a crucial role in several animal decision-making processes like movement decisions. A useful way to explore the interactions between movement decisions and learning of the spatial environment is by comparing individual behaviors during the different phases of natal dispersal (when individuals move across more or less unknown habitats) with movements and choices of breeders (who repeatedly move within fixed home ranges), that is, by comparing behaviors between individuals who are still acquiring information vs. individuals with a more complete knowledge of their surroundings. When analyzing movement patterns of eagle owls, Bubo bubo, belonging to three status classes (floaters wandering across unknown environments, floaters already settled in temporary settlement areas, and territory owners with a well-established home range), we found that: (1) wandering individuals move faster than when established in a more stable or fixed settlement area, traveling larger and straighter paths with longer move steps; and (2) when floaters settle in a permanent area, then they show movement behavior similar to territory owners. Thus, movement patterns show a transition from exploratory strategies, when animals have incomplete environmental information, to a more familiar way to exploit their activity areas as they get to know the environment better.

A stochastic representation of environmental uncertainty and its coupling to acoustic wave propagation in ocean waveguides

It is argued that a quantitative measure of incomplete environmental knowledge or information (i.e., environmental uncertainty) should be included in any simulation-based predictions linked to acoustic wave propagation. A method is then proposed to incorporate environmental uncertainty directly into the computation of acoustic wave propagation in ocean waveguides. In this regard, polynomial chaos expansions are chosen to represent uncertainty in both the environment and acoustic field. The sound-speed distribution and acoustic field are therefore generalized to stochastic processes, where uncertainty in the field is interpreted in terms of its statistical moments. Starting from the narrow angle parabolic approximation, a set of coupled differential equations is derived in which the coupling term links incomplete environmental information to the corresponding uncertainty in the acoustic field. Propagation of both the field and its uncertainty in an isospeed waveguide is considered as an example, where the sound speed is described by a random variable. The first two moments of the field are computed explicitly and compared to those obtained from independent Monte Carlo solution of the conventional (deterministic) parabolic equation that describes the acoustic wave properties.

Some issues concerning numerical simulation as a predictive tool for assessing sonar performance in uncertain environments

One goal of numerical simulation in ocean acoustics is the accurate prediction of acoustic field properties in complex environments. In this context, the objective assessment of both sonar performance and associated signal-processing schemes operating within uncertain littoral environments pose significant challenges for numerical simulation. Some of the issues involved with using numerical models for the prediction of sonar performance in the presence of incomplete environmental information are addressed. To achieve credibility as a prediction tool, the simulation results should be held to the same standards as those associated with experimental results, i.e., the numerical predictions should include error bars or confidence intervals that represent a quantitative measure of the environmental uncertainty that is always present. Providing a rational, quantitative assessment of system uncertainty in complex environments is a nontrivial task. Incomplete environmental information occurs at multiple space-time scales within both the oceanographic and acoustic modeling phases; this uncertainty can propagate, interact, and amplify in the course of transmitting and processing the signal. It is argued that environmental uncertainty should be directly incorporated into the numerical simulation process so that it links directly to the physics of wave propagation. [Work supported by NRL Base funding through ONR.]

Building maps from incomplete environment information: a cognitive approach based on the generation of expectations

Unfortunately, the real world is not crystal clear to agents. This makesmapping a hard and complex process. This paper describes a cognitive approach for mapping that relies heavily onthe generation of assumptions/expectations for the missing observational information. In addition, we present the architecture for an explorer-agent that uses this approach to build maps and whose behaviour is guided by the emotions, drives and other motivations that it may “feel”. We describe an experiment conducted in simulated environments in order to evaluate our approach for mapping

Information-Based Relative Consumption Effects

Preferences exhibit relative consumption effects if a person's satisfaction with their own consumption appears to depend upon how much others are consuming. This paper examines a model of an evolutionary environment in which Nature optimally builds relative consumption effects into preferences in order to compensate for incomplete environmental information.

Intelligent organization for flexible manufacturing

Intelligent production scheduling has been viewed as a problem-solving task that involves the generation of a suitable schedule among different ones using learning techniques. This work presents a unified model for a class of intelligent machines, suitable for flexible manufacturing systems, in which each alternative production plan comprises a fixed number of sequential steps, and the task is to select the optimal plan as more experience is obtained during the operation of the system. This model, based on the theory of hierarchically intelligent control systems developed by Saridis (1995), combines the powerful high-level decision making with advanced mathematical modeling and synthesis techniques of system theory and methods of dealing with imprecise or incomplete environment information.