A network based model for traffic sensor location with implications on O/D matrix estimates

In this paper, we define and solve the sensor location problem (SLP), that is, we look for the minimum number and location of counting points in order to infer all traffic flows in a transport network. We set up a couple of greedy heuristics that find lower and upper bounds on the number of sensors for a set of randomly generated networks. We prove that solving the SLP implies that the Origin/Destination (O/D) matrix estimation error be always bounded. With respect to alternative sensor location strategies, simulation experiments show that: (i) measurement costs being equal, the O/D estimation error is lower, and (ii) conversely, O/D estimation error being equal, the number of sensors is smaller.