VRLT – Exploring Training Measurement and Evaluation with the British Army

Brigadier Bobby Walton-Knight
Army Head of Training Capability

In late December 2018, The British Army awarded a contract to Bohemia Interactive Simulations (BISim), to demonstrate the art of the possible using virtual reality, machine learning and cloud computing for the Army’s Collective Training Transformation Programme (CTTP).

CTTP seeks to transform collective training, delivering simulation and measurement technology that prepares the Army for the dynamic and complex future operating environments.

Cervus were subsequently subcontracted to illustrate the potential benefits of enhanced data capture, data exploitation and machine learning-driven analytics within the VRLT Pilot.

01. Problem

There is recognition that the Army needs to better instrument collective training to capture, analyse and provide better feedback faster.  The employment of Hive in VRLT begins to evaluate novel TME data capture and analysis tools and in turn, inform the emerging CONEMP.

The Defence Innovation Unit funded the VRLT Pilot as part of efforts to disrupt traditional acquisition systems. Army HQ staff worked directly with industry partners to integrate in service technologies (i.e. Defence Virtual Simulation (DVS)) with Commercial off the Shelf (COTS) and innovative products. The resulting systems were then put into the hands of the user. 

02. Approach

A process of spiral development was employed to stimulate the innovation process. A baseline Virtual Reality (VR) system was developed  which enabled the participants to be trained and process/technical derisking to occur. In the following sprints, we developed and honed TME delivery and explored a variety of innovations.

For more detail on Hive please follow the link here

7.4 Million data points were captured and analysed

03. Relevance

VRLT enabled us to demonstrate:

Metrics and Data Collection

  • We developed and employed a suite of performance metrics which were derived from the existing Collective Competency Objectives to provide a logical ‘golden thread’ between training objectives and analytical outputs.
  • We successfully demonstrated the use of new and emerging performance measures to assess the development of collective knowledge skills and attitudes (KSA). We integrated data from  novel collection and analysis systems  into the overall KSA collection schema to develop more objective and non-intrusive measures to report on team performance.
  • We developed and tested the integration of Hive to DVS via a DIS interface. Where there were collection gaps we worked with BiS to develop specific DVS Applied Programming Interfaces (APIs)

Data Storage and Analysis

  • We loaded all the data onto cloud storage (to replicate G-Cloud) and then hosted our analytics engine to automate analysis.
  • We examined Hive’s ability to conduct Tactical Communications Information Flow Analysis – experimenting with automated voice to text transcription and Machine Learning to codify communications activities and exchanges.

  • We attempted “Ghosting techniques” which are prevalent in sports team analytics. We did this to investigate the use of constructive models to produce simulated outcomes against which training audience performance can be compared, given the same plan and terrain.

Reporting

  • We developed Functions in Combat and KSA dashboards to support the development of new After-Action Review processes

04. Conclusion

This was an excellent opportunity to work with the user, and to work with best in class Industry Partners, on a range of simulation and training challenges. A requirement to work under pressure of time, with rapid development iterations between each tactical event, was highly effective in getting the teams involved to innovate and stretch boundaries.

It also allowed the CTTP team to explore, test and demonstrate many of the assumptions which will eventually inform the forthcoming CTTP Concept of Employment (CONEMP).

VRLT Coverage in Soldier Magazine