With its potential to impact almost every corner of society, I can’t imagine a single sector of the economy that doesn’t stand to gain from the widespread adoption of various Artificial Intelligence, or AI, technologies. But in the military environment, it is both risky and expensive to test out new applications of AI and apply them to strategies for missions in a live military logistics supply network (MLSN).
So, can AI simulations that are currently used by the video gaming industry help the Military to improve their logistics network and decision making?
MLSNs have an array of challenges, some which are comparable to commercial logistics networks, such as balancing resources and time-constrained decision making. But that’s probably where the similarities end.
Military and commercial logistics networks differ greatly both in the degree and frequency of dynamic shifts that may occur from start to end. This dynamism causes a high level of uncertainty to an already complex field of study. Additionally, decision makers operating in the military environment are often under duress or crisis while bearing the burden that the decisions they make can ultimately cost lives.
I’ve recently been involved in a fascinating collaboration that has brought together industry and academia. Fujitsu has acted as the technical lead with this collaboration with the University of Manchester that has looked in detail at how the application of AI-based simulations currently used by the video gaming industry can help to evaluate decisions common to a military logistics supply network. The results of the project were presented recently at the 2020 NATO Modelling & Simulation Group Symposium (MSSG).
Digital twin environment
Throughout the project, we have applied computer-generated solutions to select MLSN missions. This enabled MLSN scenarios to be recreated in an artificial world, producing a potential framework for a live environment. Various scenarios could then be tested in simulation by recreating supply chain data in a digital twin environment.
By employing current gaming technologies, decision science, and simulation techniques we were able to show in great detail:
- how the computer solves the mission, action by action
- why it eliminates scenarios and
- what options are feasible.
In doing so, unit commanders can utilize the concept of reinforcement learning in day-to-day missions. They can also carry out training objectives with potential coalition partners providing visualization, decision support, what-if analysis, or problem simplification by eliminating scenarios that do not achieve the overall mission goals or violate constraints. Resolution of network issues can also be solved quickly by the use of AI, enabling a more agile, human-machine teaming approach.
Data Visualisation Observatory
A key aspect of the collaboration was having access to the Data Visualisation Observatory at Manchester University. The first of its kind in the UK, this large-scale visualisation facility offers interactive 3D visualisation of data with laser-sharp definition on a cylindrical matrix of 72 full HD screens.
The facility powers data-driven and observational research, engagement and teaching activities at the university. It will help to develop new insights into data and models, and to support information-intensive brainstorming sessions within the lab or across the world.
The possible uses for the new observatory are endless, whether it be providing high-fidelity interactive visualisations of very large data sets, capturing user behaviour in controlled immersive environments, or digital simulations of manufacturing processes and designs.
While our collaboration utilised the facility for a military specific scenario, its potential applications are far reaching, including:
- Industry 4.0: simulating innovative mechanisms for flexible routing of materials through a factory with a complex layout.
- Health: visualising health data and immersive observations of brain signals and anatomical models.
- Environment: linking air quality simulations and real-time traffic flows to reduce pollution levels.
- Marketing: observation of customer behaviour in a simulated retail environment under controlled conditions.
- Fintech: visualising monetary flows and stock market behaviours.
Find out more…
This collaboration between the University of Manchester and Fujitsu will be of great benefit to Defence logisticians, and the learnings will be taken forward into future solution development.
In terms of our future plans, we will continue this collaboration in the area of ‘Agent based modelling and decision making’, to assist our customers’ understanding within complex and uncertain operating environments. Within Fujitsu, we are continuing to strengthen our AI capabilities in areas such as predictive maintenance, NLP, computer vision and knowledge management, in order to assist our customer’s specific digital transformation journeys.
|About the Author|
|Dr. Darminder Ghataoura has over 15 years’ experience in the design and development of AI systems and services across the UK Public and Defence sectors as well as UK and international commercial businesses. Darminder currently leads Fujitsu’s offerings and capabilities in AI and Data Science within the Defence and National Security space, acting as Technical Design Authority with responsibility for shaping proposals and development of integrated AI solutions. He also manages the strategic technical AI relationships with partners and UK government.
Darminder holds an Engineering Doctorate (EngD) in Autonomous Military Sensor Networks for Surveillance Applications, from University College London (UCL).
Latest posts by Darminder Ghataoura (see all)
- What can the Military learn from video gaming applications of AI? - October 22, 2020
- Tackling the AI data challenge – could synthetic data be the answer? - July 7, 2020
- The rise of AI and the risks of relinquishing responsibility - January 21, 2020