Yu-Hong Liu

A fuzzy-neural inference network for ship collision avoidance

The basic structure of a fuzzy-neural inference network model for ship collision avoidance in sight one another is presented in this article. The model has three subnets. There are the subset of classifying ship encounter situations and collision avoidance actions, the subset of calculating membership function of speed ratio, and the subset of inferring alteration magnitude and action time. The weight values of former two subsets are obtained by self-learning from a number of samples, while those of last one subset are obtained form experience. All of these weight values can be adjusted respectively and conveniently according to practical needs. The test results show that by the inference of the model, some valuable decisions can be made from initial input data.

The structure design of an intelligent decision support system for navigation collision avoidance

A synergic multi-base structure of an intelligent decision support system for navigation collision avoidance is presented in this article. Each base, including rule knowledge base, fuzzy case base, object oriented model base and rule text base, is established and discussed. The multi-base cooperator can control, coordinate, schedule and communicate among these bases. Harmony knowledge base, dynamic database and multiple inference engines, which consist of the cooperator, are also explained. And finally, several pictures about system running and interface are given.

Case Learning Base on Evaluation System for Vessel Collision Avoidance

A learning method by recording cases based on an evaluation system for vessel collision avoidance is presented in this paper. The frame model is selected and defined as the case representation method. The basic structure and several processing modules of the evaluation system are discussed. Data fusion methods are used in the system. During processing of the learning case, fuzzy method is also adopted. And finally, the validity of the evaluation system and the learning method is proved by application instance

A CBR-Based Approach for Ship Collision Avoidance

In this paper, we propose a novel CBR-based approach for ship collision avoidance. After the introduction of the CBR-based decision-making support, we present two abstraction principles, selecting view points and describing granularity, to create collision avoidance cases from real-time navigation data. Several issues related case creation and CBR-based decision-making support are discussed in details, including case presentation, case retrieval and case learning. Some experimental results show the usefulness and applicability of CBR-based approach for ship collision avoidance.

A multiagent-based simulation system for ship collision avoidance

This paper presents a multiagent-based simulation system for the decision-making research of ship collision avoidance. The system has the characteristics of flexible agent, variable topology, isomorphic function structure, distributed knowledge storage, and integrated control method. The architecture is proposed with four kinds of agent models, that is Control_Agent, Union_Agent, Ship_Agent and VTS_Agent. We developed these agent models for modeling the behaviors for human, ship and VTS using a BDI (Beliefs, Desires, and Intentions) agent framework. The agent communication mechanism based on AIS (Automatic Identification System) message is also established and discussed. The proposed multiagent-based simulation system provides a useful platform for studying multi-target encountering problems and different decision-making methods for collision avoidance.

A multi-agent information fusion model for ship collision avoidance

A decision making model for ship collision avoidance is developed in this paper by use of the information fusion methods. The model consists of two kinds of agents, ship agent and VTS agent, and three fusion levels, original data fusion level, multiple union fusion level and distributed plan fusion level. With this model, a final collision avoidance plan with less inconsistent and more economic than distributed plans for ships within same negotiation union can be obtained.

A Case learning model for ship collision avoidance based on automatic text analysis

The sailor operation experiences are quite important for ship collision avoidance, and some of which can be found in typical collision avoidance cases. In order to use these cases effectively, it is necessary to analysis these recorded cases and learn some knowledge from them, furthermore, provide effective support for automatic collision avoidance decision making system. A case learning model based on automatic text analysis is proposed in this paper. Some useful cases and knowledge can be created from text format cases and stored in computer by use this case learning model. Some main treatments and algorithms, such as automatic Chinese word segmentation, disambiguation and semantic analysis, are discussed in this paper.

Case Learning in CBR-Based Agent Systems for Ship Collision Avoidance

With the rapid development of case-based reasoning (CBR) techniques, CBR has been widely applied to real-world applications such as agent-based systems for ship collision avoidance. A successful CBR-based system relies on a high-quality case base. Automated case creation technique is highly demanded. In this paper, we propose an automated case learning method for CBR-based agent systems. Building on techniques from CBR and natural language processing, we developed a method for learning cases from maritime affair records. After reviewing the developed agent-based systems for ship collision avoidance, we present the proposed framework and the experiments conducted in case generation. The experimental results show the usefulness and applicability of case learning approach for generating cases from the historic maritime affair records.

Multi-agent Planning for Ship Collision Avoidance

Multi-agent based planning techniques have been widely applied to various decision-making support applications. In this paper, we investigate how to apply multi-agent planning to collision avoidance in ship navigation. We have developed three multi-agent-based planning algorithms: the independent planning for self-benefit purpose, the centralized planning for union-benefit purpose and the negotiation-based planning for mutual-benefit purpose. Having introduced collision avoidance planning, we present the developed planning algorithm in detail. We also report the experiments and some results. The experimental results illustrate the feasibility and validity of the multi-agent planning for collision avoidance.

Application of Multi-agent Technology in Decision-making System for Vessel Automatic Anti-collision

By defining each vessel as an Agent, a decisionmaking system for automatic anti-collision is presented based on the theory and technology of multi-agent system. The hybrid architecture of Agent and its descriptions of Backus-Naur Form based on the Class Framework of Visual C++ are illustrated. The decision-making mechanism of the system is analyzed and some points for future research are also discussed. The problem- solving method using multi-agent organization can reduce the difficulty of solving the problem of multi-vessel collision avoidance. Rational decisions of global benefits for vessel collision avoidance can be reached by communication, cooperation and negotiation among vessels. The method was a new attempt to address the problem of multi-vessel collision avoidance, as well as to exploit a new application field for the theory of multi-agent system.