|
Home | Contact Us | Support |
|
|
|
Discovery Machine Maritime Console 1.6Discovery Machine, Inc. (DMI) has made it possible to create highly intelligent entities for use within the Joint-Semi Automated Forces (JSAF) Simulation. Discovery Machine has advanced the current JSAF capabilities in many ways and made it much easier to understand why entities are behaving the way they are. Discovery Machine’s ApproachDiscovery Machine behaviors are represented visually which allows for increased instructor understanding. The visual aspect is consistent throughout the entire behavior development process. While intelligent behaviors are being created they are represented in a hierarchical form with component names represented in English, not computer code. This allows subject matter experts (SME) to create, view, and validate behavior models prior to their use in training events. Since the behaviors automate entities in an understandable way that is validated by SMEs, instructors can spend less time driving entities and more time instructing their students. In addition, increased automated entities reduce the amount of personnel required to run a training event. This allows for larger training scenarios and enhances the training experience for trainees. Behavior Console and Training Preparation Discovery Machine has worked extensively with subject matter experts in various areas of maritime warfare. In each case DMI utilized their proven methodology to document the knowledge of those experts. Ultimately, Discovery Machine created a set of basic level actions (BLA) which represent event-specific compilations of knowledge. The BLAs range from tasks as simple as turning on sensors to elaborate maneuvers such as Searching for an Enemy Submarine. These consoles currently contain 55 BLAs for surface ships, 63 BLAs for subsurface vehicles, and 39 BLAs for helicopters. Using Discovery Machine consoles, instructors are able to leverage the expertise of top experts using pre-built BLAs. The console is designed in a way that walks the instructor through the assembly process. It organizes the BLAs based on their purpose so that an instructor can rapidly assess which BLAs to include in a mission. This allows instructors to construct a mission very quickly. At this point the instructor can review the mission and publish it directly to JSAF by the simple click of a button. This behavior is then available to assign in JSAF within a few moments without having to restart the simulation.
Resulting behaviors are fully automated and can be used to control entities in the simulation. Currently Discovery Machine has three behavior consoles; one for submarines, one for helicopter, and one for surface ship behaviors. Future plans include consoles for fixed wing aircraft, ground vehicle, and human behaviors. Highlights
Enhanced Intelligence with Discovery Machine Behaviors Discovery Machine has introduced understandable artificial intelligence to simulated entities in JSAF. This is accomplished through the entire behavior building process from authoring to deployment. All behaviors that are created are inherently intelligent because they compile the experiences of the best SMEs available.
All Discovery Machine behaviors for JSAF have three concurrent processes; situation awareness, the main mission, and sensor monitoring. The three processes are performed simultaneously which allows for intelligent behaviors that have a primary purpose but can still react to the situation at and changes to their surroundings.
Examples of intelligence in Discovery Machine behaviors can be seen through communication between both live and synthetic crews, nuanced behaviors that utilize their surroundings to decrease their chance of detection, and coordinated movement behaviors assigned to entire battle groups. In addition, all behaviors log all events down to the millisecond so that they can be used for later creation of an after action review. Highlights
Discovery Machine Integration to JSAF Discovery Machine is currently in the process of becoming a part of JSAF Baseline 4. Once integrated, Discovery Machine will be a part of subsequent official releases to the JSAF Baseline. Discovery Machine has access to the classified versions of JSAF for development and testing purposes. Discovery Machine has integrated with various software systems from other companies. These integrations have been completed to enhance the training experience for each trainee. Future integrations that will enhance the end user training experience are being considered. Highlights
|
Anyone who has a basic understanding of the JSAF simulation is capable of assigning Discovery Machine behaviors to characters within the simulation. The intelligence in the behaviors is understandable because of the way it is represented. It is represented visually in a decision tree format so that anyone who uses Discovery Machine behaviors can see what an entity is doing and understand why it is doing it.
Discovery Machine’s concept of situation awareness has grown out of Mica Ensley’s cognitive research. Each entity controlled with a Discovery Machine behavior is able to process raw sensory input to create a sense of itself and its surroundings. The situation awareness processor enables increased comprehension of the simulated “world” so that it can interact with it intelligently. Equipped with the information provided by the situation awareness processor, entities can respond to the situation intelligently, react to new information, and alter their mission accordingly.