In one concept of embedded training, the goal is to outfit a military vehicle with computer hardware capable of battlefield simulation, and use this to train personnel in the field. The advantages are two-fold. First, personnel are trained on the same equipment they use in real-life exercises. Second, personnel can be trained locally; there is no need to transport them to dedicated simulation facilities that house expensive, single-use machinery. Such a training environment could benefit from the use of Computer Generated Forces (CGF) to provide automated opponents against which personnel may train. However, a critical problem with the use of CGF in embedded training is space: current CGF systems run on relatively large workstations that will not fit into operational equipment.

This paper discusses the feasibility of one technical approach for effective embedded training using CGF systems: hardware-based CGF. Rather than using an unwieldy and expensive high-performance workstation to provide the CGF functionality, it may be possible to construct a "CGF processor" using Very Large Scale Integration technology. A CGF processor would provide comparable performance at a size much smaller than the general-purpose workstations currently being used for CGF applications. This miniaturized "CGF-on-a-chip" would fit into preexisting onboard computer systems in operational equipment, eliminating the need for bulky workstations to provide CGF.

This paper identifies the set of operations performed by a CGF system, including those operations that are particularly relevant to embedded training, and are thus good candidates for implementation in hardware. The paper illustrates these ideas concretely by discussing a hardware algorithm for one CGF operation.