Tomasz Praczyk

Evolving Co-Adapted Subcomponents in Assembler Encoding

Evolving Co-Adapted Subcomponents in Assembler Encoding The paper presents a new Artificial Neural Network (ANN) encoding method called Assembler Encoding (AE). It assumes that the ANN is encoded in the form of a program (Assembler Encoding Program, AEP) of a linear organization and of a structure similar to the structure of a simple assembler program. The task of the AEP is to create a Connectivity Matrix (CM) which can be transformed into the ANN of any architecture. To create AEPs, and in consequence ANNs, genetic algorithms (GAs) are used. In addition to the outline of AE, the paper also presents a new AEP encoding method, i.e., the method used to represent the AEP in the form of a chromosome or a set of chromosomes. The proposed method assumes the evolution of individual components of AEPs, i.e., operations and data, in separate populations. To test the method, experiments in two areas were carried out, i.e., in optimization and in a predator-prey problem. In the first case, the task of AE was to create matrices which constituted a solution to the optimization problem. In the second case, AE was responsible for constructing neural controllers used to control artificial predators whose task was to capture a fast-moving prey.

Decision system for a team of autonomous underwater vehicles-Preliminary report

The paper presents preliminary research whose main goal is to build a ship protection system. One of the elements of the system will be a sub-system responsible for capturing and destroying/neutralizing all dangerous objects occurring inside a guarded area. This task will be performed by a team of autonomous underwater vehicles. To construct a decision system for the vehicles, i.e. the system whose task is to provide high-level decisions regarding direction and velocity of move, many different methods can be applied. Examples are evolutionary neural networks and expert systems with fuzzy decision rules. To test usefulness of the above methods in a role of the decision system, experiments in a predator-prey problem were carried out. In the experiments, the task of vehicles-predators was to capture a vehicle-prey behaving by a simple deterministic strategy. The current paper is a report on all the experiments mentioned.

Neural anti-collision system for Autonomous Surface Vehicle

Autonomous Surface Vehicles (ASV) are robots destined for the operation at water basins like lakes, canals, harbors and even open sea. They are used for different purposes, e.g. for patrol tasks, as scouts, or as a support for Navy ships. One of the main tasks of ASV is to move along a fixed path to a destination point. To this end, an anti-collision system (ACS) has to be used with the ability to lead ASV along a path and to simultaneously avoid all additional objects present at the basin, e.g. ships, sailing boats, fishing cutters, icebergs. To perform this task, the ACS implemented as an evolutionary neural network can be used. The paper describes architecture of the neural ACS, presents two neuro-evolutionary methods used to build the system, and reports the whole process of constructing it.

Using augmenting modular neural networks to evolve neuro-controllers for a team of underwater vehicles

The paper presents a new generative neuro-evolutionary method called augmenting modular neural networks (AMNN). As the name of the method implies, its purpose is to construct neural networks with a modular architecture. In addition to the modularity itself, neural networks evolving according to AMNN are also characterized by gradually expanding architecture. In the beginning of the evolutionary process, all networks consist of only output modules (or a single module). After some time, if the architecture of all networks is insufficient to effectively perform a task, all of them are augmented by one hidden module. In the following generations, further hidden modules are also added and this procedure is continued until some stopping criterion is satisfied. To test performance of AMNN, the method was used to evolve neuro-controllers for a team of underwater vehicles whose common goal was to capture other vehicle behaving by a deterministic strategy (predator–prey problem). The experiments were carried out in simulation, whereas their results were used to compare AMNN with neuro-evolutionary methods designed for building monolithic neural networks.

Solving the pole balancing problem by means of assembler encoding

The neuro-evolution is a domain of artificial intelligence which uses the evolutionary approach to produce artificial neural networks. There are many neuro-evolutionary methods and one of them is Assembler Encoding. This paper compares Assembler Encoding with other methods from the range of neuro-evolution and and reinforcement learning. During comparison tests, the task was to form neuro-controllers for three variants of the inverted pendulum problem. The variants differed in the amount of information supplied to each neuro-controller and in the number of poles installed on a cart.

Software architecture of biomimetic underwater vehicle

Autonomous underwater vehicles are vehicles that are entirely or partly independent of human decisions. In order to obtain operational independence, the vehicles have to be equipped with a specialized software. The main task of the software is to move the vehicle along a trajectory with collision avoidance. Moreover, the software has also to manage different devices installed on the vehicle board, e.g. to start and stop cameras, sonars etc. In addition to the software embedded on the vehicle board, the software responsible for managing the vehicle by the operator is also necessary. Its task is to define mission of the vehicle, to start, to stop the mission, to send emergency commands, to monitor vehicle parameters, and to control the vehicle in remotely operated mode. An important objective of the software is also to support development and tests of other software components. To this end, a simulation environment is necessary, i.e. simulation model of the vehicle and all its key devices, the model of the sea environment, and the software to visualize behavior of the vehicle. The paper presents architecture of the software designed for biomimetic autonomous underwater vehicle (BAUV) that is being constructed within the framework of the scientific project financed by Polish National Center of Research and Development.

Research on biomimetic underwater vehicles for underwater ISR

Autonomous Biomimetic Underwater Vehicles BUVs driven by an undulating propulsion are a new branch in an area of an underwater robotics. They imitate both the construction and kinematics of a motion of underwater living organisms, e.g. fishes. Such vehicles have several features crucial from the point of view of military applications, e.g. larger secrecy and potential range of operation. The paper presents results of the research on BUVs carried out within two (Polish and EDA) projects both led by Polish Naval Academy. At the beginning, the initial efforts in building Polish BUV called CyberFish are included. Then, selected results of the tests of subsystems, e.g. navigational and 3D model of BUV built within national project are described. Next, the initial research achieved in the international project are showed. At the end, the schedule of the research planned to carry out within both projects is inserted. The paper is mainly focused on the hardware development of the BUVs.

Using evolutionary neural networks to predict spatial orientation of a ship

The ability to precisely predict behavior of a ship can be useful for different ship systems, for example, dynamic positioning, video or artillery systems. To automatically maintain a ship position and heading, the dynamic positioning system has to know future behavior of the ship as exactly as possible. The same applies to the ship video and artillery systems which to continuously track a target have to predict both the location of the target and orientation of the ship deck in relation to the target in successive points in time.In this paper, evolutionary neural networks are proposed as ship behavior predictors. To perform the task, they are supplied with the information about ship spatial orientation (Euler angles) acquired from inertial navigational systems. In experiments reported in the paper, both monolithic and modular recurrent neural networks were tested. To build them, a neuro-evolutionary method called Assembler Encoding with Evolvable Operations was applied. As the point of reference for the networks two other prediction methods were used: the first is Linear Regression with Correction, i.e. the most effective method in preliminary experiments, whereas the second is Autoregressive Integrated Moving Average, which seems to be currently the most general tool for forecasting a time series.

Application of Bearing and Distance Trees to the Identification of Landmarks on the Coast

Application of Bearing and Distance Trees to the Identification of Landmarks on the Coast The problem of continuous position availability is one of the most important issues connected with the human activity at sea. Because the availability of satellite navigational systems can be limited in some cases, e.g. during military operations, one has to consider additional methods of acquiring information about the ships position. In this paper one of these methods is presented, which is based on exploiting landmarks located on a coastline. A navigational radar is used to obtain information about these points. In order to estimate the ships position by means of a set of landmarks, it is necessary to know their accurate locations. The paper presents a landmark identification method based on the comparison of bearing and distance trees representing pattern points generated from a chart, as well as points extracted from a radar image.

The Influence of Parameters of Biomimetic Underwater Vehicle Control System on the Ability of the Vehicle to Avoid Obstacles

Autonomous underwater vehicles are vehicles that are entirely or partly independent of human decisions. In order to obtain operational independence, the vehicles have to be equipped with a specialized software that usually has many different parameters. The parameters decide about effectiveness of the software and in consequence the vehicle. The paper reports experiments performed in simulation, whose goal was to analyze the influence of selected parameters of High-level Control System of Biomimetic Underwater Vehicle on the vehicle ability to avoid obstacles.