Robert A. Ghanea-Hercock

Co-operating mobile agents for distributed parallel processing

1.1 The Team Agent Approach This paper describes a novel approach to dynamic distributed parallel processing using a mobile agent-based infrastructure. Our goal is to extend the concept of the Parallel Virtual Machine architecture [8] by using a combination of collaborative and mobile software agents to enable automatic and dynamic configuration of distributed processes. The remainder of this paper describes a mobile agent based DVM that involves several agent roles, with different functions and skills. Our approach to distributed processing is regulated by a twotier management system. At the strategic level, an anchored centralised agent is responsible for managing user interaction and determining how tasks are to be distributed. Whilst the mobile agents who deliver code to remote machines, manage local processing at an operational level. We therefore developed specialised mobile agents, each performing particular roles, which co-operate as a team to achieve user defined goals. The resulting system provides users with a much simpler means of utilising the power of distributed parallel computing. As these agents behave co-operatively as a team our system is called MATS: the Mobile Agent Team system. One of the design objectives of MATS was to minimise the code that would need to be moved about the network. This was achieved by localising the most complex functionality in an anchored (non-mobile) agent, which can then communicate with the simpler specialised mobile agents through a message passing mechanism. In effect, this is a marriage of the autonomous and social features of ‘collaborative’ agents with the facilities offered by mobile agent technology.

Many hands make light work? An investigation into behaviourally controlled co-operant autonomous mobile robots

The past ten years has seen a urry of research activity into the behavioural control of autonomous mobile robots. Yet despite this effort, many researchers are of the opinion that behavioural robots are incapable of achieving tasks more complex than simple can collecting, box pushing, herding or moving in formation. If such robots are to gain industrial credibility, these criticisms must be addressed. To focus the research we have studied the application of multiple mobile robots to a complex nuclear plant decommissioning problem. We argue that it is possible for multiple mobile robots to co-operatively perform a complex task provided that solutions to a number of key issues are incorporated into a behavioural control architecture. These include: behaviour con ict resolution, behaviour adaptation and behaviour scheduling. We have designed behavioural control methods to address these issues and our work has resulted in the creation of a behaviour synthesis architecture (BSA) which has been implemented on two real mobile robots. The application of the BSA to our complex industrial task is detailed and the results from the work are presented.

Disturbed behavior in co-operating autonomous robot

Balancing the conflicting demands imposed by a dynamic world on an autonomous robot requires a significant degree of adaptability. This paper describes a multi-layer control system for two co-operating mobile robots, which uses fuzzy logic to adapt the relative importance of a set of reactive behaviours. The fuzzy system therefore acts as an arbiter, which smoothly interpolates control between conflicting behaviours. This allows the robots to successfully navigate out of local potential minima. The adaptive mechanism itself is also modified by an array of vectors generated from an on-line analysis of the activity of each fuzzy rule. From recent work on neural dynamics [Kelso, 171 the strategy is to consider the control system as a dynamic structure, and to achieve adaptivity through maintaining it in a disturbed or stressed phase condition. This is achieved by monitoring the matrix of fuzzy rules, and triggering a suppression of rules which are driving the system into a stable state. We propose that for an autonomous agent in an unstructured environment maintaining a state of dynamic instability within the control system increases the probability of the agent reaching its goal.

What Does a Planner Need to Know About Execution

This paper discusses how far the characteristics of the execution systems impact the planner which plans for them. It does this in the context of the MACTA project in which multiple cooperating mobile robots running a behavioural architecture are integrated with a Reflective Agent based on a computer running a symbolic planner, in this case UCPOP. The paper proposes some solutions but concludes that it is very difficult to produce a clean interface between a planner and its execution systems.

PLANNING AND BEHAVIOURS - A HYBRID ARCHITECTURE FOR MOBILE ROBOTS

Abstract The paper discusses a hybrid predicitve-behavioural architecture being developed by the Mobile Robotics Research Group at Salford University for cooperating mobile robots. The test bed used for experiments is described, the multi-agent framework for the hybrid architecture is explained and the behavioural and predictive components are presented. Issues of division of responsibility, communication nd the relationship between the two styles of architecture are described.