Space Engineers by Keen Software House is a sandbox game where players find resources and construct ships, buildings, or sculptures out of ‘blocks’. The interaction space between the players and tools and blocks is huge, so exhaustive testing is prohibitively expensive.
The iv4XR framework allows developers and testers to improve the testing process by creating and including agents who either have a specific goal and script to follow, or are given a more open ended objective where they can explore the interaction space and find potential error states.
Making use of the iv4XR agents has reduced the amount of manual testing required for the game. Automating some of the most rote tests have freed the testing team to investigate more difficult bugs.
Maev by Thales is a simulation environment that is used to simulate large complex systems like the infrastructure of a city or the movement of crowds. One of the uses for Maev is in simulating security infrastructure for critical sites. In iv4XR the pilot application is a scenario in which there is a nuclear power plant that requires security systems (cameras, guards, motion sensors, etc). A human operator is to place these systems around the scenario and then intruders controlled via iv4XR will attempt to breach the placed security.
Previously, human operators were used in all parts of the development pipeline – both the placing of security systems and intrusion were handled by different teams. Iv4XR allows the intrusion to be automated which allows for a tighter feedback loop in designing security for critical infrastructure.
Iv4XR has allowed the testing loop for the nuclear power plant intrusion simulation to be administered by only a single person instead of the team that was required previously.
LIVEsite by Gameware is a monitoring system for construction projects, from buildings to civilian infrastructure. To monitor the integrity of these projects, a number of sensors are integrated into the construction itself. These sensors monitor some aspect of the building (tensility, pressure, load, etc) and report this data. Depending on the poll rate of the sensor, this data can be in the form of thousands of datapoints per second.
In such a large volume of data, potential issues can be obscured. While there are basic thresholds that can be configured for detecting outliers, this data has to be flagged for an engineer to look over.
Iv4XR alleviates this by having its agents probe errant readings more thoroughly. Given the readings and knowledge of the sensor, an agent can create tests which can help determine the causes of alerts and recommend courses of action to the engineers based on the readings and subsequent tests.
Using iv4XR in LIVEsite has already detected intermittent faults in critical rail infrastructure that had hitherto been hidden by the volume of data from the installed sensors.