José Soler

Modelling Agents in Hard Real-Time Environments

Over the last few years more complex techniques have been used to develop hard real-time systems, and the multi-agent system paradigm seems to be an appropriate approach to be applied in this area. The temporal restrictions of these systems made necessary to build agents architectures that satisfy these restrictions. A formal Agent architecture to model hard-real time systems is proposed in this paper. A prototype that follows this agent architecture and works in a hard real-time environment has been implemented. Finally, a study case is described to show the design agent process.

Applying the ARTIS Agent Architecture to Mobile Robot Control

The agent/multi-agent system paradigm is an important field of Artificial Intelligence. The use of this paradigm in real-world problems is one of the main lines of interest in this area. To do this, it is necessary to make use of agent/multi-agent architectures and artefacts. This paper describes an architecture for real-time agents, an artefact (InSiDE) for the specification of agent-based systems and the application of this architecture for the control of an autonomous mobile robot.

SIMBA: An Approach for Real-Time Multi-agent Systems

The use of the agent/multi-agent system paradigm has increased sharply as an important field of research within the Artificial Intelligence area. In recent times, the application of this paradigm seems appropriate for solving complex problems which require intelligence and bounded response times. This paper presents Simba : an architecture based on ARTIS agents as its main component for the development of real-time multi-agent systems. The ARTIS agent architecture guarantees an agent response that satisfies all its critical temporal restrictions in a real-time environment. The main feature of SIMBA systems is their applicability for complex, distributed, real-time domains. The architecture allows the communication among agents taking into account their hard temporal restrictions. n nIn order to show the use of systems of this kind, the paper describes the design of a multi-agent system for the distributed intelligent control of a residential building.

On the combination of high-pressure and low-pressure exhaust gas recirculation loops for improved fuel economy and reduced emissions in high-speed direct-injection engines

In this paper, an experimental study of the combination of low-pressure and high-pressure exhaust gas recirculation architectures has been carried out. In the first part of the paper, the effects of both high-pressure and low-pressure exhaust gas recirculation architectures on engine behaviour and performance are analysed by means of a series of steady tests. In the second part, the effects of the combination of both architectures are addressed. The results show that the low-pressure configuration improves high-pressure exhaust gas recirculation results in brake-specific fuel consumption, nitrogen oxides and exhaust gas opacity; nevertheless, hydrocarbon emissions are increased, especially during the engine warm up. In addition, the exhaust gas recirculation rate achieved with low-pressure systems is limited by the pressure difference between diesel particulate matter outlet and compressor inlet; therefore, the high-pressure system can be used to achieve the required exhaust gas recirculation levels without increasing pumping losses. In this sense, the combination of both exhaust gas recirculation layouts offers significant advantages to reduce emissions and fuel consumption to meet future emission requirements.

Potential of Using a Nozzle at the Compressor Inlet of a High-Speed Direct-Injection Diesel Engine

The present work reports the effect of a nozzle placed upstream of the compressor of a high-speed direct-injection diesel engine. It has been observed that mounting a nozzle at the compressor inlet can lead to a shift in the compressor surge line to lower mass flowrates. In order to study the effect of the nozzle on the engine performance a set of experimental investigations has been conducted including characterization of the nozzle alone using a flow test rig up to sonic conditions and characterization of the nozzle—compressor unit by determining its operation map in a turbocharger test bench. The distortions in the compressor inlet are usually considered as a disadvantage, as they affect the efficiency and the behaviour of the turbocharger. However, the comparison between the original compressor map (provided by the manufacturer) and the nozzle—compressor map given in the present paper shows that the change in the flow pattern caused by the nozzle improves the engine—turbocharger coupling and thus reduces the air mass flowrate, leading to a surge by 30 per cent, for a given pressure ratio. The improvement in turbocharging does not necessarily increase the performance of the engine. Nevertheless, the possibility of increasing the boost pressure allows elevated fuel flowrates to be applied while still maintaining the combustion efficiency, thus increasing the power of the engine

Social conformity and its convergence for reinforcement learning

A dynamic environment whose behavior may change in time presents a challenge that agents located there will have to solve. Changes in an environment e.g. a market, can be quite drastic: from changing the dependencies of some products to add new actions to build new products. The agents working in this environment would have to be ready to embrace this changes to improve their performance which otherwise would be diminished. Also, they should try to cooperate or compete against others, when appropriated, to reach their goals faster than in an individual fashion, showing an always desirable emergent behavior. In this paper a reinforcement learning method proposal, guided by social interaction between agents, is presented. The proposal aims to show that adaptation is performed independently by the society, without explicitly reporting that changes have occurred by a central authority, or even by trying to recognize those changes.

Multi-Agent systems over RT-Java for a mobile robot control

There are a great number of programming languages and frameworks for multi-agent system development, but most of them do not allows us create and later manage real-time multi-agent systems. In this paper, we propose the use of the RT-Java language for real-time multi-agent development. We also present the jART platform as an appropriate framework for the execution of agents of this kind. A mobile robot control is implemented and analyzed using this framework.

Developing a Mobile Robot Control Application with CommonKADS-RT

The advances achieved in Real-Time Artificial Intelligence (RTAI) allow us to apply techniques to build large applications. However, developing such applications require having a methodology that defines the techniques, tools and processes to allow for modeling their characteristics. There are several important lines of research in RTAI, such as architectures, languages, methods and methodologies. All of them provide the necessary information and knowledge to be able to build computer systems based on real-world problems. This paper describes CommonKADS-RT, a method to develop Real Time Intelligent Systems. This method is based on CommonKADS with the addition of necessary elements to model real time restrictions. We also present an application of this approach to the control of an autonomous mobile robot.

Social Reinforcement Learning for Changing Environments

If we imagine a dynamic environment whose behavior may change in time we can figure out the difficulties that agents located there will have trying to solve problems related to this environment. Changes in an environment e.g. a market, can be quite drastic: from changing the dependencies of some products to add new actions to build new products. The agents should try to cooperate or compete against others, when appropriated, to reach their goals faster than in an individual fashion, showing an always desirable emergent behavior. In this paper a reinforcement learning method proposal, guided by social interaction between agents, is presented. The proposal aims to show that adaptation is performed independently by the society where which these AI-controlled players belong, without explicitly reporting that changes have occurred by a central authority, or even by trying to recognize those changes.