Gamasutra - Features - "AI Middleware: Getting into Character, Conclusion"

After some 16 years of designing and developing AI solutions for many different genres of computer games, I was admittedly skeptical that commercial AI middleware could help the game development process in any significant way. After all, I thought, the AI for every game is unique, and thus requires a craftsman such as myself to design and develop good AI solutions. While that might be true to some degree, this week's survey of AI middleware revealed to me that these tools offer viable game development solutions.

The four previous installments of this week's series focused on the capabilities of each product. Today's concluding installment summarizes these products and offers some thoughts on where you should go next. Admittedly, because this is just a summary, the article does not do justice to these products. So I encourage you to check the web sites of these products and learn more about their specifics.

AI.implant is a sophisticated animation and character control engine from BioGraphic Technologies in Montreal, Canada. The most unique aspect of AI.implant is the use of Maya and 3DS Max plug-ins to allow AI to be added to the character during the modeling process. Once a character has been created in Maya/3ds max, the AI.implant plug-in can be used to set animation control for the character, add attributes to the character, set default and initial state values, add sensors and assign behaviors to the character and create decision trees that manage the character's behavior.

Characters can be grouped and coordinated using AI.implant, and default behaviors can be assigned in lieu of complex decision-making. Decisions are made using binary decision trees; a system that supports sophisticated rule-based processing. AI.implant also offers waypoint editing (via the plug-ins) and automatic waypoint-network generation.

DirectIA (which stands for "Direct Intelligent Adaptation") was developed by Mathematiques Appliquees S.A. of Paris, France. DirectIA relies on several built-in engines for processing:

DirectIA offers real-time decision and action behavior modeling tools. Within the high-level tools is support for complex agents and reactive agents. The low-level tools offer pathbuilding, hierarchical pathfinding and steering tools (which are currently in beta).

DirectIA is very agent-centric. Its sophisticated behavior engine relies on its own built-in functionality tuned via script files to meet the needs of your agent's decision-making system. Agent communication and perception can be integrated via user-defined callback routines accessed from the scripts. The tools decision-making process is very "state" oriented, and should appeal to developers who favor the use of finite state machines.

RenderWare AI Middleware (RWAI) was developed by Kynogon, a French company specializing in game AI (the Kynogon Artificial Intelligence Modules, or KAIM), and is packaged as a part of the RenderWare Platform suite of game development tools marketed by Criterion Software of Austin, Texas. The entire RenderWare suite includes modules for handling graphics, audio and physics, but I only looked at the module for AI.

The RWAI SDK focuses on designing and implementing character behavior. RWAI views the objects in the world as entities, in two basic forms: thinking entities (which contain a Brain object) and passive entities (objects that exist in the world and interact with thinking entities). RWAI provides several layers of services:

RWAI is a powerful AI Middleware SDK to the game developer, and when used in combination with the other RenderWare Platform components, is a very complete solution for game developers.

SimBionic is a visually oriented tool for defining characters, implementing their decision-making capabilities, and assigning their behaviors. It was developed by Stottler Henke of San Mateo, California.

Like DirectIA, SimBionic is very state-oriented. This means that most control flow is influenced by the state or condition of some object or process. The SimBionic state systems have many components that can be classified as descriptors and declarations:

SimBionic provides a sophisticated framework for creating and debugging state systems. The SimBionic Visual Editor makes developing these state systems easy and accessible to level designers and other staff who may not have deep programming skills. Since FSMs are so widely used in game development, SimBionic could be an alternative to custom FSM development.

In his article "Effective Middleware Evaluation" in the May, 2003 issue of Game Developer magazine, Alex Macris made several valid points about the process of evaluating middleware as a solution to computer game development issues. They were:

As more games tout AI as a selling point, and if the "not invented here" syndrome wanes and more third-party middleware is embraced by developers, then I believe AI middleware can become an important solution for game developers.

	Product
Feature	AI.implant	DirectIA	RWAI	SimBionic
User Decision Support	Binary decision trees	Motivated decision graphs	Finite state machines, neural networks	Finite state machines
Other Services	Auto path generation, pathfinding	Pathfinding	Graphics and physics in other modules, auto path generation, pathfinding	Inter-agent communication
Behavior Support	Pre-packaged behaviors	Templated behavior scripts	Pre-packaged behaviors	User developed behaviors
Engine Source Code Availability	Some	No	Yes	Some
Extensibility	User-developed behaviors	User-developed scripts, callback functions	User-developed behaviors, callback functions	User developed behaviors, callback functions
Production Tools	Maya/3ds max plug-ins	Script templates, tuning GUI	AI skeleton code, XML configuration	Visual editor, visual debugger

Table 1. Summary of common features of reviewed AI Middleware Products.

______________________________________________________