Cogntive Agents in Game Environment (Demo)

"I call you squishy and you shall be mine."[Finding Nemo]

The following game demo is largely based on Unity's own Stealth demo.

I changed the robots a bit by controlling them through my cognitive architecture POSH# and augmenting the selection mechanism with ERGo. The main reason is that the resulting behaviour is less deterministic in terms of strict behaviour patterns.

The robots have five drives: attack,charge,chase,patrol, and idle. Charge was added in contrast to the stealth demo to make the demo more interesting as the robots now have conflicting motives. If they do not re-charge frequently they run out of juice and are stuck, unable to attack or move. So when a player is visible they might or might not follow him depending on their internal motivation. Same applies to patrolling and attacking. The new dynamics are currently really rough but the additional overhead of including POSH# and ERGo for 5 agents is really low which is super. The biggest impact on performance in this demo is the use of Unity's NavMeshAgent and its path recalculation.

Demo requires webgl to work in your browser