Study of mobile mixed sensing networks in an automotive context

Mixed sensing mobile networks comprise of mobile sensors that have different sensing capabilities. We look at such sensor networks in an automotive context; wherein automobiles with two levels of sensing (and consequently with two different dynamics) are ‘mixed’ among one another. The two levels of sensing considered are local, near-neighbor information sensing; and advance, far-ahead information sensing. We look for conditions governing the way the two types of sensors should be mixed (i.e., required minimum number and distribution of the far-ahead information sensing vehicles in a mixed N-vehicle string) in order to meet certain performance objectives. In this regard, two types of models are considered – microscopic models (using ODEs) governing individual vehicle behavior; and macroscopic models (using PDEs) governing average behavior of groups of vehicles. The performance objective that we address is related to the safety of the overall network, and depends on the type of model being adopted – thus in the microscopic model, the performance metric is one of achieving zero collisions, in conditions where there otherwise would have been multi-vehicle collisions; while in the macroscopic model, the metric is one of weakening the shock waves that otherwise would have existed.