摘要：LVC 训练是实战化条件下装备体系对抗训练的一种有效手段，针对 LVC 训练系统中，计算机生成兵力难以满足训练需求问题，明确 LVC 训练与 LVC 训练系统概念，按照模型与系统的结构组成关系阐述了逻辑靶场实体配置、指挥实体、战斗实体 3 个不同层次模型相应的建模技术需求。针对具体需求，提出基于复杂网络的逻辑靶场虚实实体配置、基于深度强化学习的分队战术决策建

模、基于动态贝叶斯网、遗传神经网络的战术行为参数矫正建模 4 种计算机生成兵力生成技术。

Computer generated forces are simulated entities that are used in simulation based training and decision support in the military. The behaviour of these simulated entities should be as realistic as possible, so that the lessons learned while simu-lating are applicable in real situations. However, it is time consuming and diÿcult to build behaviour models manually, and there has been an increasing interest in automating this process using machine learning.

The military battlespace is often visualized as set of layers representing different aspects, ranging

from physical terrain to information flows. Computer Generated Forces (CGF) simulations used for

campaign and mission simulation have traditionally focused on the physical representation of units,

更多

terrain and effects. In 2016 a high level simulation architecture was proposed that integrates

concepts of information flow, content and cognition with the traditional physical layer. The aim of

that work was to provide a common architecture to guide the evolution of CGF as modern warfare

environments and effects such as cyber are required to be added.

收起

ABSTRACT: Rules of Engagement (ROE) are driven by a mix of legal, military, and political factors. These dimensions can interact and overlap in subtle ways and must be carefully crafted to be easy to apply in combat situations without jeopardizing mission outcome and the warfighter’s right to self-defense. Although trial and error may have sufficed in the past, the growing complexity of conflicts and the military and political ramifications of ineffective ROE (e.g., a friendly fire incident), make a simulation-based ROE evaluation system a high priority. This paper describes ROE3, a human behavior-modeling tool that supports tactics-independent representation of ROE. In our approach, ROE are defined as meta-knowledge that act as a constraint on the tactical choices selected by the synthetic entity. This is key to the flexibility of the system — tactics and ROE can be freely mixed and matched to investigate their interactions.

ABSTRACT

Computer Generated Forces (CGFs) are a key component in constructive simulations and are being increasingly used to control multiple entities in Synthetic Environments (SEs). Being a cost-effective way to providing extra players in SEs, they are becoming a possible alternative in various activities, such as Concept, Development and Experimentation (CD&E), analysis, training, tactic development, and mission rehearsal. The predictable nature of many current CGFs behaviour is one of their biggest problems, making it easy for the trainee to distinguish between human-controlled and computer-controlled entities in the simulation environment. This can result in negative or ineffective training as the trainee quickly learns to predict the behaviour of the CGF entity and easily defeats it in a way that would not happen with a human opponent. This results in a requirement for humans to control synthetic entities, thus limiting simulation exercises by the availability of operators. If instead the Artificial Intelligence (AI) of these entities could be improved, the number of operators required will, thus, be reduced. The first step in such an effort is evaluating the AI capabilities commonly available in CGFs. Such an analysis was performed at the Defence Research & Development Canada (DRDC), revealing the common strengths and weaknesses of available CGFs, and suggesting which might be most useful as a platform for further AI research. This document presents the methods and results of this analysis.

Distributed Interactive Simulation (DIS) is an architecture for building large -scale

simulation models from a set of independent simulator nodes communicating via a

common network protocol. DIS is most often used to create a simulated battlefield for

更多

military training. Computer Generated Forces (CGF) systems control large numbers of

autonomous battlefield entities in a DIS simulation using computer equipment and

software rather than humans in simulators. CGF entities serve as both enemy forces

and supplemental friendly forces in a DIS exercise. Research into various aspects of

CGF systems is ongoing. Several CGF systems have been implemented.

收起

Computer generated forces (CGFs) are autonomous or semi-autonomous actors within military, simulation

based, training and decision support applications. The CGF is often used to replace human role-players in

military exercises to, ultimately, improve training eff ciency. The modeling and development of CGFs is a

更多

complex, time-consuming and expensive endeavor where military domain expertise and doctrinal knowledge

are interpreted and programmed into the CGF by hand. Furthermore, CGFs often represent human actors

and behaviors (pilots, soldiers, manned systems, etc.) making it an even more challenging task.

收起

Abstract—Commercial/Military-Off-The-Shelf (COTS/MOTS) Computer Generated Forces (CGF) packages are widely used in modeling and simulation for training purposes. Conventional CGF packages often include artificial intelligence (AI) interfaces, but lack behavior generation and other adaptive capabilities. We believe Machine Learning (ML) techniques can be beneficial to the behavior modeling process, yet such techniques seem to be underused and perhaps under-appreciated. This paper aims at bridging the gap between users in academia and the military/industry at a high level when it comes to ML and AI. We address specific requirements and desired capabilities for applying machine learning to CGF behavior modeling applica-tions. The paper is based on the work of the NATO Research Task Group IST-121 RTG-060 Machine Learning Techniques for Autonomous Computer Generated Entities.