Luís Correia

A Useful Autonomous Vehicle with a Hierarchical Behavior Control

This paper describes a vehicle to perform a simple task using a decision control composed of a behavior hierarchy. First the model is presented. It is based on two fundamental components: the behavior module defining how each behavior element is implemented; and the blocker which is the basic component of the arbitration system, responsible for behavior selection. Behavior modules are composed of two submodules, one to produce actuator actions and other to produce priorities associated to actions. They also have three parameters whose values can be changed prior to execution, as a means of configuring a particular control architecture. Competition among behaviors is resolved by arbitration structures, each one consisting of a series of blockers working in a completely distributed way. The arbitration system can be automatically generated from the mere enumeration of the constituting behaviors. The development of the model in order to support behavior hierarchies is also presented as well as the possibility of using global variables. Then the prototype — a vehicle to clear away film boxes — is described in detail namely all the behaviors constituting its control. Finally some conclusions based on the obtained results are presented.

Behavior-Based Architecture with Distributed Selection

A behavior based architecture is presented in which behaviors are sparsely interconnected, using an arbitration structure which solves the problem of behavior selection in a distributed way. The arbitration structure is constituted by sets of blockers, each blocker resolving conflicts between two behaviors. There will be a blocker set for each actuator in order to allow independent actions from non interfering behaviors. The behavior model is also described comprising two sub-modules, one for the action and another for the priority output of the behavior. The user when designing a particular vehicle must also define, for each behavior, a fixed priority input plus a fatigue and a recovery times that complete the behavior characterization. A case study is depicted with analysis of the vehicle global behavior for different parameter values. The implications of the architecture model for the control of autonomous vehicles are discussed and future developments are advanced

Adapting populations of agents

We control a population of interacting software agents. The agents have a strategy, and receive a payoff for executing that strategy. Unsuccessful agents become extinct. We investigate the repercussions of maintaining a diversity of agents. There is often no economic rationale for this. If maintaining diversity is to be successful, i.e. without lowering too much the payoff for the non-endangered strategies, it has to go on forever, because the non-endangered strategies still get a good payoff, so that they continue to thrive, and continue to endanger the endangered strategies. This is not sustainable if the number of endangered ones is of the same order as the number of non-endangered ones. We also discuss niches, islands. Finally, we combine learning as adaptation of individual agents with learning via selection in a population.

Fuzzy Behaviors and Behavior Arbitration in Autonomous Vehicles

In this paper a fuzzy behavior based decision control is presented. A behavior is described by one or more fuzzy controllers reacting to external stimulus, and a state variable (denoted behavior activity) measuring how evident are the conditions of application of the knowledge coded in the fuzzy controllers. The autonomous vehicle decision control is composed by several behaviors of this kind, each one feeding an arbitrator with the activity and action values. The arbitrator evaluates the final action to send to the actuators taking into consideration each behavior activity and action values. Two arbitration methods are proposed. Results of simulations are presented and conclusions are drawn.

A Resource Sharing Model to Study Social Behaviours

We analyse a resource sharing game under different conditions in order to study social behaviours such as cooperation and treason. We concentrate in analysing the game under different scenarios in order to find out which one produces a more cooperative population. This is possible since the game while having multiple Pareto Optimal strategies, also has multiple pure and mixed Nash Equilibriums, which results in complex population dynamics. We introduce a simple form of agreement and compare the results with the previous cases.

Expressing Population Based Optimization Heuristics Using PLATO

There are presently many and seemingly different optimization algorithms, based on unrelated paradigms. Although some nice and important intuitions support those heuristics, there is (to our knowledge) no rigorous and systematic approach on how to relate them. Herein we present a framework to encompass those heuristics, based on the multiset formalism, providing a common working structure and a basis for their comparison. We show how to express some well known heuristics in our framework and we present some results on relations among them.