In the present paper, single-vehicle data of highway traffic are analyzed i
n great detail. By using the single-vehicle data directly, empirical time h
eadway distributions and speed-distance relations can be established. Both
quantities yield relevant information about the microscopic states. Several
fundamental diagrams are also presented, which are based on time-averaged
quantities and compared with earlier empirical investigations. In the remai
ning part, time-series analyses of the averaged as well as the single-vehic
le data are carried out. The results will be used in order to propose objec
tive criteria for an identification of the different traffic states, e.g.,
synchronized traffic. [S1063-651X(99)06712-4].