|Acquisition Program: || Objective: ||The objective of this effort is to research, develop and implement prototype capabilities to simulate the communication and coordination of an operational Intelligence Surveillance and Reconnaissance environment for the purpose of providing: (a) Realistic RPA crew mission task saturation; (b) enhanced integration and coordination across an operational testbed environment; (c) a system to prototype, integrate and evaluate intelligent agents and synthetic teammates from both government and commercial sources at various levels of fidelity.
|| Description: ||In August 2010 operational Remotely Piloted Aircraft (RPA) crews visited AFRL Mesa to attend a training effectiveness workshop. During the workshop operators were asked for ideas to improve training. The consistent answer was to provide a capability for the RPA operators to interact and coordinate with other members of the ISR community (e.g. Mission Intelligence Coordinators) as they do in real world operations. Again in September 2010 during a follow up meeting with additional RPA crews, the same question was posed and the same answer was given. A system is needed that will create realistic task saturation in a simulated environment for ISR communication and coordination for improved training effectiveness for RPA operations and other command and control assets. This system should employ synthetic and intelligent agent technology to provide operationally realistic communication and coordination in the midst of a scenario but also have the capability to monitor individual performance and adapt the level of saturation to provide the appropriate level based on individual operator capabilities, essentially adaptive training for personalized learning. Additionally, the system should be designed to utilize existing operational community text based communication mediums (e.g. mIRC chat) to communicate. The modeled entities should be constructed synthetically and formatted so that both government and COTS systems may be used in parallel. The level of fidelity can interchangeably range from basic scripting to advanced intelligent architecture. The system should be capable of supporting multiple simulators for command and control training within an operational testbed.
|| ||PHASE I: Will research the operational needs, and result in the development of the initial underlying software architecture for this system and demonstrate its capabilities in a proof of concept to show functionality of each component of the total system.
|| ||PHASE II: Will fully develop the prototype software technology and underlying architecture and demonstrate the capability to provide task saturation and adapt the levels of saturation in an operational testbed in three different scenarios.
|| ||PHASE III DUAL-USE COMMERCIALIZATION:
Military Application: The military is increasingly relying more and more on chat communication for C2 operations. The development of this system would provide a capability that would apply to several domains and provide a new capability to improve operator ability to handle task saturation through personalized and adaptive training.
Commercial Application: This capability provides an adaptive training technology for task saturation for chat based communications in civilian contexts.
|| References: ||
(1) Cantwell, H.R., “Operators of Air Force Unmanned Aircraft Systems Breaking Paradigms” Maxwell AFB, AL, AIR & SPACE POWER JOURNAL SUMMER, 2009.
(2) Daniel Gonzales et al., Network-Centric Operations Case Study: The Stryker Brigade Combat Team, Santa Monica, Calif.: RAND, MG-267-1-OSD, 2005
|Keywords: ||Command and Control Training, Chat Communication Training, Task Saturation, Personalized Learning, Adaptive Training, Remotely Piloted Vehicle Training|
Questions and Answers:
Q: Can we assume to have two operational mode of the envisioned training simulation environment? (A) "expert" mode for the RPA expert crews where we learn good practices and behavior, and (B) "novice" mode where we provide cognitive support to enhance performance using learned good practices.
A: You can assume anything you want. If you think it is the best approach then go for it. I am looking for innovation.
Q: It seems some linguistics processing is necessary to understand the content of a chat session to provide support.
- Does the effort have scope of such processing?
- If yes, how much?
A: The level and depth (if any)of linguistic processing will be for your team to decide.
Q: 1. Will the government provide some guidance to build a realistic scenario?
2. Will they provide any chat data?
A: 1. You need to plan on receiving zero subject matter expertise from the government. That being said I will try to help wherever practical.
2. No chat data will be available.
Q: What's the level of simulation fidelity needed? In other words, how intelligent are agents simulating other crews? Assuming simple rule-based GPS-type implementation would suffice instead of fine grained cognitive models such as Soar and ACT-R.
A: This will be something your team needs to decide. Some of the players are more complicated than others.
Q: Do you have a performance measurement metric in mind to evaluate the effective of this kind of simulation environment in enhancing crews' performace?
A: I have no specific metrics in mind. Performance measurement and tracking systems that are DIS capable are very popular in our community.
Q: What is the priority of coordination between the entities? For instance, focus on the coordination between the operator and ISR community agents, or focus on the coordination between ISR community agents, or both?
A: The focus for this effort is the RPA operators and the communications they receive from the community.
Q: Should we assume that there are no chat monitoring assistance tools for the operators within the environment?
Q: 1. Are you interested in simulating/emulating multiple RPAs at the same time?
Or many command and control personnel for one RPA?
2. Does this topic focus on individual operator or multiple crew members with different roles and responsibilities?
3. Are you interested in simulating/emulating realistic channel characteristics between RPA and command & control, for example, delay, channel blockage, Interference, or fading? Also, are you interested in using real radios in emulation/simulation environment?
A: 1) Both, but it will be RPA centric.
2) This topic focuses on the Pilot and Sensor Operator.
3) I am interested in providing a task saturation environment, if you think these pieces will help provide the best environment then include them.
Q: Will we have access to the training environment(s) that are targeted for integration?
A: The teams that receive awards and have proper clearances will be permitted access on a prearranged basis.
Q: 1. Are there unclassified scenarios that will be provided to topic awardees as GFI?
2. Are there unclassified transcripts that will be provided to topic awardees as GFI?
A: 1. No.
Q: 1. How many Phase I and Phase II awards do anticipate under this topic?
2. Will this effort be competed both within and between topics?
A: 1) I can not disclose that information. Also, please limit questions to technical clarification of the topic.
2) Once again, please limit questions to technical clarification of the topic.
Q: 1) "The modeled entities should be constructed synthetically and formatted so that both government and COTS systems may be used in parallel." Are there Specific Government or COTS systems in mind here?
2) Phase I Objective "Will research operational needs" -- is there a target operations or training center where this reserach should be conducted? Will AFRL be able to support access to such a site?
A: 1) No specific systems have been identified.
2) There is no specific operations center or training center identified. AFRL may potentially be able to support a sight visit but there is no guarantee. Do not plan on it for your proposal.
Q: 1) Is there a particular targeted UAV Platform in mind?
2) Is there a minimum security clearance level required to perform this work?
A: 1) The MQ-1 and MQ-9.
2) No, this system can be built at the unclassified level. Scenarios could potently become classified.
Q: 1. This is a very ambitious and broad SBIR topic. It’d be useful to better understand what aspects of the problem are the intended primary foci.
1a. What’s the #1 problem you’d like to see addressed in Phase I? In Phase II?
1b. What user need is most critical?
1c. What does success (at the end of Phase II) look like?
1d. What is the primary goal for this SBIR?
2. Are there specific technical or deployment requirements that the training and intelligent agent technology needs to meet?
2a. Processing power requirements?
2b. Form factor requirements?
2c. Number of simultaneous agents?
2d. Number of simultaneous missions?
2e. Computer generated agent reaction time requirements?
2f. Scenario Authorability?
2g. Natural language interaction with humans?
2h. Behavior Transparency (easy for people to understand what the agents are doing and why)?
2i. Behavior that adheres to doctrine?
A: 1a. Please see the objective paragraph of the original proposal.
1b. Please see the objective paragraph of the original proposal.
1c. Please read the proposal.
1d. Please read the proposal.
2. No, be innovative.
2a. An adequate amount of processing power to meet your system processing requirements.
2b. Whatever you determine to be reasonable.
2c. Whatever you determine to be reasonable.
2d. Whatever you determine to be reasonable.
2e. Whatever you determine to be reasonable.
2g. Whatever you determine to be reasonable.
2h. Whatever you determine to be reasonable.
2i. Whatever you determine to be reasonable.
Q: 1. Is part of the training objective to address flaws or weaknesses in the electronic signal communication with the RPA?
2. Should poor signal transmission be modeled by real communication systems, or should they be modeled within the simulation.
3. Should weaknesses in electronic signals be relegated to Phase II efforts?
A: 1) No.
3) If you feel electronic signal degradation and/or weaknesses has an important role to play by all means put it into your proposal. You may want to reread the topic.