Procedural Model Generation Service Helps Improve Training

By Dolly Rairigh Glass

A new software tool, the Procedural Model Generation Services (PMGS), is directly impacting the success of the Rapid Data Generation Program of Record by providing high-quality models to external simulation systems without the need for a large hand-created library, thus enhancing the training environments to improve Soldier readiness.

The Research & Development program under which PMGS is being generated is co-funded/managed by Program Executive Office for Simulation, Training and Instrumentation (PEO STRI), and the U.S. Army Research Laboratory (ARL), Human Research and Engineering Directorate (HRED), Simulation and Training Technology (STTC).

“Our current warfare challenges demand more than ever that soldiers “train as they fight,” said Hector J. Gonzalez, Science and Technology Manager for Synthetic Natural Environments, STTC. “Therefore it is of paramount importance that we provide them with a simulated environment as realistic as possible in an expedited manner. PMGS contributes to enhance both the speed of production and the quality of our simulation environment thus improving Soldier readiness.”

The Procedural Model Generation Service (PMGS) reduces the time needed for graphic artists to integrate 3-D models into simulations, and looks to be a possible value added for many programs across Team Orlando in the future as they begin to investigate perhaps how PMGS might fit into their programs.

And because the program is provided with Government Purpose Rights, any other government program has the right to use it without incurring more costs. Other programs can use it as is free of charge or fund the modifications they need to make it work within their own simulations.

“PMGS allows graphic artists to create 3-D models, like buildings, and then utilize them, along with the terrain database, within a simulation like Combined Combat Tactical Trainer (CCTT),” said Dr. Robert Cox, Assistant Program Manager, Rapid Data Generation, PEO STRI Environmental Representation/Geospatial Lead. “PMGS reduces the time needed for the graphic artist, therefore saving on labor, time and money.”

“Currently, the time table to produce a building, for example, is about two to four hours,” explained Cox. “With PMGS, we are able to create that same building in seconds. It’s a huge cost-avoidance.”

Simulation systems like the CCTT provide several realistic virtual environments, and PMGS is more cost-efficient because it is streamlined, enabling the user to access a pre-built model library.

“There are two types of models,” Cox said, “geo-specific, a graphic artist’s depiction of a building, let’s say the White House, and a geo-typical, a graphic artist’s depiction of a typical house built in Florida. PMGS builds those geo-typical models based on a series of ‘procedures’ or ‘business rules.’”

Cox continued, “For instance, a business rule might be all geo-typical houses in Florida are made out of concrete blocks. PMGS defines the business rules that are then applied to the development of a 3-D model.”

As the program manager for the Rapid Data Generation, one of the programs that co-funded GameSim to produce the Procedural Model Generation Service, Cox was looking to develop technology for Synthetic Environment (SE) Core, a sister program.

SE Core develops environmental representation databases for PEO STRI and U.S. Army simulations. “It’s environmental because it is a confluence of terrain, models and other things,” Cox said. “A lot of people say terrain, but for us, that is not totally correct because simulations use more than terrain. Synthetic Environment Core develops these databases and they are used by virtual and constructive simulations to train the soldiers on specific tasks.”

The co-funder/manager of PMGS, the U.S. Army’s Simulation Training and Technology Center, is also extending the PMGS capability by adding procedural generation of vehicle models capability to the tools set.

“In addition to transitioning to a Program of Record, like SE Core, STTC is also using the Procedural Model Generation Service baseline in our research testbed at STTC,” said Julio de la Cruz, Chief Engineer, SNE, STTC.

“The scope of the investment involves the research and investigation on
the effects of the time and cost that PMGS has for producing terrain databases
for simulated training systems,” de la Cruz said. “There is more and more demand for 3D models in computer graphics, virtual reality, and modeling and simulation, and this results in a change in emphasis for the requirements.
The visual quality becomes one of the main points of attention.”

Modeling of three-dimensional (3D) objects sequences is a challenging problem, and has been a research topic for many years. De la Cruz said that in the database generation process, history has shown the generation of 3D Models presents a bottleneck that is categorized as an impact on cost and time during the production process.

Currently, STTC has only performed analysis for 3D building models, but future research will include broader types of models, like vehicle and live form 3D models to help them obtain better accuracy through maximum likelihood estimation.