Statechart-Based AI in Practice
نویسندگان
چکیده
Layered Statechart-based AI shows considerable promise by being a highly modular, reusable, and designer friendly approach to game AI. Here we demonstrate the viability of this approach by replicating the functionality of a full-featured and commercialscale behaviour tree AI within a non-commercial game framework. As well as demonstrating that layered Statecharts are both usable and amply expressive, our experience highlights the value of several, previously unidentified design considerations, such as sensor patterns, the necessity of subsumption, and the utility of orthogonal regions. These observations point towards simplified, higher-level AI construction techniques that can reduce the complexity of AI design and further enhance reuse.
منابع مشابه
Model-Driven Development of Game AI: Research Plan
The field of game AI is largely industry driven, lacking an agreed upon formalism for AI representation. Adhoc scripting languages, simple finite state machines, behaviour trees, and planners are employed, but not in a fashion adhering to any standard. As a result, reuse is sparse between games and formal analysis techniques are undeveloped. As research for a Ph.D. thesis, we propose to show th...
متن کاملCommunication and Synchronization of Distributed Medical Models: Design, Development, and Performance Analysis
Model-based development is a widely-used method to describe complex systems that enables the rapid prototyping. Advances in the science of distributed systems has led to the development of large scale statechart models which are distributed among multiple locations. Taking medicine for example, models of best-practice guidelines during rural ambulance transport are distributed across hospital s...
متن کاملThe ArmarX Statechart Concept: Graphical Programing of Robot Behavior
ed through the Sensor–Actor unit concept. These Sensor–Actor units realize hardware or simulator access and hide the low-level communication from higher-level layers of the robot software. On the mid-level, robot capabilities, such as perception, planning, and motion generation, are implemented in a network transparent way. The high-level layer comprises a set of robot skills, realized as state...
متن کاملModel-Based Design of Game AI
The complexity of modern computer games has increased drastically over the last decades. The need for sophisticated game AI, in particular for Non-Player Characters (NPCs) grows with the demand for realistic games. Writing meaningful, consistent, modular, re-useable and efficient AI code is not straightforward. In this article, we suggest to model rather than to code game AI. A variant of Rhaps...
متن کاملAutomatic Test Cases Generation for Statechart Specifications from Semantics to Algorithm
This paper studies automatic test cases generation for Statechart specifications. Applying Tretmans’ approach to generate test cases from labeled transition systems, we provide a solid mathematical basis for conformance testing and automatic test case generation for Statechart specifications. In order to make the test cases generation process more simple and efficient, we propose observable sem...
متن کامل