You are here

Mission Success Assessment and Mitigation Recommendations Using a Cognitive System

Description:

TECHNOLOGY AREA(S): Air Platform 

OBJECTIVE: Develop a cognitive system as a selectable Unmanned Aircraft System Control Segment (U CS) Service with potential application in Naval Air Systems Command (NAVAIR) Common Control System (CCS). 

DESCRIPTION: The demands on unmanned vehicle operators are increasing with the evolution of autonomous vehicles. As the CCS evolves, it is expected that one operator may need to manage a large number of autonomous manned and unmanned vehicles of varying capabilities and vehicle management needs. It is important that the operator knows where, how and when to place attention on needed issues during the execution of an integrated mission plan, especially where multiple vehicles are making decisions autonomously “ without operator approval, or management by negation. Because of these newly anticipated demands on the operator, there is a need to develop a cognitive system that is offered as a UCS Service that is assessing mission risk with some form of statistical confidence to help ensure mission success within this challenging environment. This cognitive system, should support the operator in knowing where, how and when to intercede with autonomous operations to ensure mission success when controlling multiple/diverse vehicles within a theater of operations, even when the theater of operations is fluid and demanding. In order to determine if a cognitive system is best in class with regard to providing a UCS service, the following is provided as the currently identified, but not necessarily all, the criteria that will be considered with regard to autonomous control systems (ACS) and CCS: 1. How well does the cognitive system candidate conform to ACS based on the real time control system architecture? 2. How difficult/easy can the cognitive system candidate be used within the UCS architecture? With regard to analyzing the cognitive systems quality of design to support best in class determination, the following is provided as the currently identified, but not all of, the criteria that will be considered: 1. From what source was the knowledge acquisition process to develop the cognition performed? a. Were multiple sources used? b. Did the sources have differences in perspective? c. If so, how were they resolved to support an optimal cognitive action? 2. What was the knowledge acquisition process used, including the process of translating expert knowledge to a cognitive system, network or branch structure? 3. With regard to the cognitive system, how was the system tested/proven to be reliable in its assessment and recommendations for mission success? 4. Is the statistical confidence associated with the cognitive system improving with increased sample size? The cognitive system is required to demonstrate reliability in terms of mission risk assessment and providing recommendations based on that assessment for unmanned vehicles. The cognitive system should assess both individual vehicle/mission risk and collaborative vehicle/mission risk, including multiple unmanned vehicles and unmanned vehicles supporting manned aircraft, for example an MQ-25s support of multiple F/A-18s. Included in the stages of developing the cognitive system, it is important that first a simulation and then live demonstration of the cognitive system be demonstrated with six or more autonomous, manned and unmanned vehicles. Of the six vehicles, an ultimate end goal for acceptance of the technology is that a minimum of two vehicle types carrying different sensor payloads need to be included in both the simulation and live demonstration. Threshold capability would be two or more cooperative unmanned systems; objective capability would be two or more cooperative manned and unmanned systems. The cognitive system should be designed as a service module that can be installed within a UCS compliant environment based SAE standard AS6518 to support the control stations three main components: (1) Vehicle Management, (2) Mission Planning, and (3) Mission Management. The cognitive system needs to store enough past flight information associated with all three components to support learning. The learning should be specifically focused on how to better determine success and assess risk from previous related and unrelated missions and platforms, without burdening overall Control Station performance. The cognitive system should utilize UCSs Data Distribution Services (DDS) middleware within the Control Station, including but not limited to sensor information to assess mission success from the vehicle. The cognitive system should also provide recommendations to the operator to improve mission success, along with percent of confidence increases or other risk assessment improvement. The recommendations should be able to be translated into UCS message commands that provide vehicle and payload management. Additionally, the cognitive system should include safety alerts and various safety or risk levels in real time and continuously during mission operations. A goal of the UCS cognitive service is to cause a ground control station (GCS) to become an autonomous station, controlling various autonomous vehicles to successfully complete mission goals under supervision of a single operator. This autonomous GCS is envisioned to integrate available sensor information from one or more unmanned autonomous vehicles under control by the GCS. The cognitive system shall be able to assess with a level of statistical confidence whether a mission plan will be successful. The assessment should include a measure of risk associated with the mission plan. Additionally, the cognitive system should provide safety alerts to the operator based on assessment of risk along with recommendations to mitigate/resolve the safety alerts. 

PHASE I: Provide a learning-based, algorithm in the form of a UCS service. The algorithm needs to be able to collect and integrate sensor information from various sources under the operators control. The algorithm should include the cognitive design of a learning system that uses statistical confidence to support success and risk assessment of a mission requiring a variety of unmanned vehicles. The algorithm should show how it can reliably and accurately determine the statistical confidence as to whether the mission plan will be successful, along with the degree of risk, including recommendations and safety alerts associated with each vehicle under the control of the GCS. 

PHASE II: Develop and demonstrate a prototype cognitive system in the form of an UCS service within a UCS compliant GCS. The cognitive system should be able to collect available sensor information from various unmanned vehicles under the GCSs control. The prototype demonstration should show how the cognitive system, using the algorithm developed in Phase I, can reliably and accurately determine the statistical confidence as to whether the mission plan will be successful and to what degree is the risk of failure, including recommendations and safety alerts to improve the degree of risk. Before a prototype is developed, a simulation should be developed to successfully show that the algorithms code has been implemented properly. It should integrate various sensor data that supports target and friendly vehicle identification tracking, mission assessment and targeting for potential kill chain solutions. Once the algorithms code is proven within the simulated environment, a live demonstration will be required, where one or more vehicles, controlled by a GCS running the algorithm, are following a complex mission plan. During the execution of the mission plan, data should be collected and processed by the algorithm in the form of one or more UCS services. The algorithm should be able to provide real time risk assessment and mitigation recommendations to the operator of the GCS. 

PHASE III: Based on a successful prototype demonstration, further develop and test the cognitive systems algorithm as demonstrated in Phase II for transition to the NAVAIR CCS program. In this phase, both a simulation and live demonstration are required that control a minimum of six vehicles following a multi-stage mission plan, where a minimum of two different vehicle types are used and the algorithm is providing real time assessments and recommendations. This technology will benefit large delivery organizations such as United Parcel Service, FedEx, and others that focus on using autonomous unmanned air vehicle delivery of parcels and other items. The ability for a cognitive system to forecast mission success of one or more vehicles, while also making mitigation recommendations, has applications throughout aviation, robotics and unmanned systems industries, including commercial applications and other ground control systems within airports. Private Sector Commercial Potential: This will benefit large delivery organizations such as United Parcel Service, FedEx, and others that focus on using autonomous umanned air vehicle delivery of parcels and other items. The ability for a cognitive system to forecast mission success of one or more vehicles, while also making mitigation recommendations, has applications throughout aviation, robotics and unmanned systems industries, including commercial applications and other ground control systems within airports. 

REFERENCES: 

1: RCS: The Real-time Control Systems Architecture (2011). Rep. NIST, n.d. Web. Specifically review sections on Development Support and Performance Measures. Retrieved from http://www.nist.gov/el/isd/rcs.cfm; Giordano, J., Wurzman, R (2016)

2: Integrative Computational and Neurocognitive Science and Technology for Intelligence Operations: Horizons of Potential Viability, Value and Opportunity. Retrieved from http://www.potomacinstitute.org/steps/featured-articles/85-integrative-computational-and-neurocognitive-science-and-technology-for-intelligence-operations-horizons-of-potential-viability-value-and-opportunity

3: Ernst, R (March 2016). UCS Architecture Overview, NAVAIR presentation. Retrieved from http://www.dtic.mil/ndia/2016GRCCE/Ernst.pdf

4: Sun R (2007). Introduction to computational cognitive modeling. Retrieved from 7 December 2016 from http://www.sts.rpi.edu/~rsun/folder-files/sun-CHCP-intro.pdf

5: Forsythe, F, Giordano, J (2011). On the Need for Neurotechnology in the National Intelligence and Defense Agenda: Scope and Trajectory. Synthesis: A Journal of Science, Technology, Ethics and Policy 2 no 1, (2011): 5-8. Retrieved from http://www.synesisjournal.com/vol2_no2_t1/Forsythe_Giordano_2011_2_1.pdf

6: Unmanned Systems (UxS) Control Segment (UCS) Architecture: UCS Architecture Model. http://www.sae.org/search/?sort=date&content-type=(%22STD%22)&root-code=(%22AS6518%22)

 

KEYWORDS: Cognitive System; UCS Architecture; Vehicle Management; Mission Management; Mission Planning; Common Control System; Sensor Information; Autonomous Vehicles; Risk Assessment 

US Flag An Official Website of the United States Government