Sensor Maximization

Sensors

Regardless of how they are defined, spider will attempt to maximize an arbitrary number of sensors. For the first versions of spider, the only recorded sensor is the moving average of the speed at which spider is traveling. To maximize this sensor, spider must record both its decisions and pertinent environmental information and relate it to its speed. This information can then be modeled in a concise format and referenced when spider needs to make a decision.

Sensor Abstraction

#TODO A sufficiently complex model of control features, environmental features, and sensors may benefit from the abstraction of these sensors. In essence, some high level sensors are defined as favorable, but models are built that use the favorability of other sensors as input, instead of direct control features. The spider then understands which subsensors to maximize in which situations, in order to maximize the high level sensors. This system requires multiple experiential models, but is able to exclude the consideration of low level control features when making decisions about high level problems.