In this paper, a series of system-theoretical issues in the frame of a traffic systems engineering are comprehensively addressed and discussed with a particular focus on a nonlinear dynamics perspective. Transportation systems are complex (systems of systems) nonlinear dynamical systems and their comprehensive engineering does involve amongst others the following steps: modeling/identification, simulation, systemstate observation/measurement, and control/optimization. The particular message of this paper is that a significant improvement can be reached on each of these fronts, while compared to most related traditional approaches, by involving a series of adapted/optimized methodological instruments inspired and/or taken from the broad area of nonlinear dynamics, which also includes neuro-computing. For each of the above-cited steps, related limitations of most traditional approaches are presented. To finish, a selection of own novel contributions on each of the steps are shortly presented for illustrative purposes.