AI Primary: GeckoSuper
Each GeckoSavant excels at its particular task, but they can achieve very little unless they work together. GeckoSuper is what allows this coordination to take place. It not only passes information betweeen other GeckoSavants, it also monitors that information to decide if additional intervention is required.
- Coordinates all GeckoSavants
- Can make high-level decisions that individual GeckoSavants cannot
- Extensible, can support more GeckoSavants
GeckoSuper lies at the heart of the Mobile Service Robot's (MSR) software architecture. It acts as the central hub through which all other GeckoSavants communicate with each other. As this occurs, GeckoSuper can monitor all of the data flow and make actionable, "common sense" decisions about that data. For example, the CareBot could currently be reading a bedtime story, and if an onboard sensor detects smoke, carbon monoxide or another hazard, GeckoSuper will tell the CareBot to stop what it's currently doing and alert and coordinate the other GeckoSavants in such a way that the overall behavior results in leading the care receiver to safety. If the other GeckoSavants represent the CareBot's lower brain functions, it could be said that GeckoSuper represents the higher brain functions.
GeckoSuper's design is also flexible and extensible enough to allow the inclusion of future GeckoSavants and other software.
GeckoSuper is an AI/expert system responsible for system-wide orchestrated subsumption. At its simplest, subsumption means organizing various simple behaviors into layers. Low-level layers contain the most basic behaviors (eg. avoid an obstacle), while high-level layers can be thought of as more abstract behaviors (eg explore the surroundings). This is a bottom-up approach; low-level layers can override the high-level layers as necessary. The end result is an approach where high-level layers provide an overall goal for the AI system to achieve whereas the low-level layers provide the means.
This approach summarizes GeckoSuper perfectly. Its ability to receive data from other GeckoSavants and make decisions based upon that data means that GeckoSuper can provide a general directive for the robot to currently execute without having to actually worry about how it is carried out. This frees up GeckoSuper to instead focus on prioritization; given two or more inputs, GeckoSuper can determine the order in which they need to be addressed. Depending upon the circumstances, this could be interpreted as a form of common sense machine intelligence.