Scaling-up behaviours in EvoTanks: Applying subsumption principles to artificial neural networks

Thompson, Tommy; Levine, John

doi:10.1109/cig.2008.5035635

Cited by 7 publications

(16 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Each time a higher layer is added, the complexity of the task is incremented [14]. Thompson and Levine recently used a similar method to develop a layered controller for the EvoTanks game [15].…”

Section: B Learning Hierarchical Controllersmentioning

confidence: 99%

Hierarchical controller learning in a First-Person Shooter

Hoorn

Togelius

Schmidhuber

2009

2009 IEEE Symposium on Computational Intelligence and Games

View full text Add to dashboard Cite

Abstract-We describe the architecture of a hierarchical learning-based controller for bots in the First-Person Shooter (FPS) game Unreal Tournament 2004. The controller is inspired by the subsumption architecture commonly used in behaviourbased robotics. A behaviour selector decides which of three sub-controllers gets to control the bot at each time step. Each controller is implemented as a recurrent neural network, and trained with artificial evolution to perform respectively combat, exploration and path following. The behaviour selector is trained with a multiobjective evolutionary algorithm to achieve an effective balancing of the lower-level behaviours. We argue that FPS games provide good environments for studying the learning of complex behaviours, and that the methods proposed here can help developing interesting opponents for games.

show abstract

Section: B Learning Hierarchical Controllersmentioning

confidence: 99%

Hierarchical controller learning in a First-Person Shooter

Hoorn

Togelius

Schmidhuber

2009

2009 IEEE Symposium on Computational Intelligence and Games

View full text Add to dashboard Cite

show abstract

“…The SNA controllers (shown in Figure 3) were devised in our previous research found in [12] and are inspired by the seminal works by Brooks in [13], [14]. Each controller is composed of a series of ANNs existing in a hierarchy of precedence.…”

Section: Controller Librarymentioning

confidence: 99%

“…Allowing us to create unique behaviours through simple combinations. Please see [12] for further information on the controller construction and training methodology.…”

Section: Controller Librarymentioning

confidence: 99%

“…This provides variety in execution since the behaviour will not necessarily be the same in each instance. Each controller is pre-trained and ready to use straight away, given the time available for development and testing we borrowed resulting chromosomes from our experiments in the EvoTanks domain [1], [12] to test our architecture. While there are differences in the environments, our results in Section V show that these were still sufficient for our testing purposes.…”

Section: Controller Librarymentioning

confidence: 99%

See 1 more Smart Citation

Realtime execution of automated plans using evolutionary robotics

Thompson

Levine

2009

2009 IEEE Symposium on Computational Intelligence and Games

Self Cite

View full text Add to dashboard Cite

Abstract-Applying neural networks to generate robust agent controllers is now a seasoned practice, with time needed only to isolate particulars of domain and execution. However we are often constrained to local problems due to an agents inability to reason in an abstract manner. While there are suitable approaches for abstract reasoning and search, there is often the issues that arise in using offline processes in real-time situations. In this paper we explore the feasibility of creating a decentralised architecture that combines these approaches. The approach in this paper explores utilising a classical automated planner that interfaces with a library of neural network actuators through the use of a Prolog rule base. We explore the validity of solving a variety of goals with and without additional hostile entities as well as added uncertainty in the the world. The end results providing a goal-driven agent that adapts to situations and reacts accordingly.

show abstract

“…Approaches to divide the tasks such that the main behaviours emerge as a collection of small tasks has been envisioned through the evolution of subsumption like architectures e.g. [2,28,31,35,54] as well as the incremental evolutionary approach [1,17,56] and more recently the layered learning approach [51,55]. All these techniques involve the subdivision of tasks into various modules to solve the problem and the revision of the fitness function.…”

Section: Introductionmentioning

confidence: 99%

Evolving robot sub-behaviour modules using Gene Expression Programming

Mwaura

Keedwell

2014

Genet Program Evolvable Mach

View full text Add to dashboard Cite

Many approaches to AI in robotics use a multi-layered approach to determine levels of behaviour from basic operations to goal-directed behaviour, the most well-known of which is the subsumption architecture. In this paper, the performances of the unigenic gene expression programming (ugGEP) and multigenic GEP (mgGEP) in evolving robot controllers for a wall following robot is analysed. Additionally, the paper introduces Regulatory Multigenic Gene Expression Programming (RMGEP), a new evolutionary technique that can be utilised to automatically evolve modularity in robot behaviour. The proposed technique extends the mgGEP algorithm, by incorporating a regulatory gene as part of the GEP chromosome. The regulatory gene, just as in systems biology, determines which of the genes in the chromosome to express and therefore how the controller solves the problem. In the initial experiments, the proposed algorithm is implemented for a robot wall following problem and the results compared to that of ugGEP and mgGEP. In addition to the wall following behaviour, a robot foraging behaviour is implemented with the aim of investigating whether the position of a specific module (sub-expression tree (ET)) in the overall ET is of importance when coding for a problem.

show abstract

Scaling-up behaviours in EvoTanks: Applying subsumption principles to artificial neural networks

Cited by 7 publications

References 10 publications

Hierarchical controller learning in a First-Person Shooter

Hierarchical controller learning in a First-Person Shooter

Realtime execution of automated plans using evolutionary robotics

Evolving robot sub-behaviour modules using Gene Expression Programming

Contact Info

Product

Resources

About