89 Transportation Research Record: Journal of the Transportation Research Board, No. 2315, Transportation Research Board of the National Academies, Washington, D.C., 2012, pp. 89–99. DOI: 10.3141/2315-10 B. Ciuffo, Institute for the Environment and Sustainability, and V. Punzo, Institute for Energy and Transport, Joint Research Centre, European Commission, Via E. Fermi, 2749–21027 Ispra (VA), Italy. M. Montanino, Department of Transporta- tion Engineering, Università di Napoli Federico II, Via Claudio, 21–80125 Napoli (NA), Italy. Corresponding author: B. Ciuffo, bciuffo@unina.it. used in the design of devices working onboard the vehicles to assist drivers to maintain safe and comfortable driving conditions (e.g., intelligent speed adaptation systems, and collision avoidance systems). The high number of car-following models proposed could be motivated by their overall incapability to reproduce both traffic prop- agation and driver–vehicle interactions without relying on the over- fitting produced by their parameters, which are usually unnecessary and without a clear physical interpretation (this difficulty certainly poses serious concerns about the models’ capability to reproduce unpredictable conditions). As a result, most of the applications using car-following models usually adopt the less recent, “classical” models. The Gipps’ car-following model is one of them. Reasons for the fascination with the Gipps’ model primarily reside in the clear physical context that the author adopted to derive it: a driver adapts his speed in order to smoothly reach the desired speed or to safely proceed behind his leader. In addition, Wilson (5) demon- strated that, similar to other “reductionist models” like that of Bando et al. (6), the Gipps’ model may also allow for a uniform flow of traffic to lose stability for certain ranges of its parameters. Stability loss is an important feature since it allows for typical traffic mechanisms to be produced (such as flow breakdown and spontaneous traffic jam formation). However, as noted by Ranjitkar et al. (7 ) and Spyropoulou (8), some properties of the Gipps’ model have been hidden by the posi- tions assumed by Gipps himself and thus the scope of the model might even be enlarged. For all these reasons, in this study the main features of the Gipps’ car-following model as they have been derived in different studies and applications are summarized. Furthermore, some analyses are presented that were performed on the acceleration component of the Gipps’ model, providing some insight on the effect that the relax- ation of three parameters, usually considered as fixed, may have on the overall model performance. At the end of the analysis different versions of the Gipps’ model as derived from the literature review and the analyses carried out will be compared. The comparison is made possible by calibrating and validating the different versions in order for them to reproduce vehicle trajectory data. ORIGINAL FORMULATION The Gipps’ car-following model is the most commonly used model pertaining to the class of “safety distance” or “collision avoidance” models. Models of this class aim to specify a safe following distance and to adapt the driver’s behavior in order to always keep the dis- tance. The basic idea behind the model is that each driver plans his or her speed for the following instant (i.e., after a delay τ) such that he or she can safely stop even in the event of the leading vehicle’s sudden Thirty Years of Gipps’ Car-Following Model Applications, Developments, and New Features Biagio Ciuffo, Vincenzo Punzo, and Marcello Montanino Researchers and practitioners commonly use car-following models for road traffic studies. Although dozens of models have been presented so far, the one proposed by Peter G. Gipps in 1981 is still one of the most extensively used. However, many features of the model are still not well known or neglected in common applications. In this context, the current study summarizes and analyzes the main findings available in the scien- tific literature for the Gipps’ car-following model and introduces some of its novel features that may improve its capability to reproduce real trajec- tory data. In particular, the structure of the acceleration component of the model is analytically investigated for what concerns the meaning of some parameters that in common practice are usually kept to some empirically derived fixed values. Possible versions of Gipps’ model are presented, and their performance to reproduce real vehicle trajectories is evaluated and compared. The results achieved show the necessity for these parameters to be calibrated to improve the model’s predictive capabilities. In 1981, P. G. Gipps published a study titled “A Behavioural Car- Following Model for Computer Simulation” (1). This paper was bound to have a considerable impact on traffic flow theory and practice, and the model described became widely recognized as the Gipps’ car-following model. Car-following models try to explicitly reproduce the complex dynamics governing the actions of the driver–vehicle system while the driver is following another vehicle. Dozens of car-following models had been presented previously and new ones are continuously being proposed; since an up-to-date comprehensive review is not available, the reader can refer to various sources presenting a clustered review of the topic (2–4). Different assumptions have been presented on the strategy adopted by a driver–vehicle system to adapt its speed to the presence of a vehicle downstream moving in the same direction. Car-following models have two main applications: modeling macroscopic traffic propagation and evolution and modeling the microscopic behavior of the driver–vehicle system during follow- the-leader activity. In the first case, car-following models are usually included within a broader modeling framework of traffic micro- simulation (acting as the main driver for the vehicles’ movements) and the main aim is to simulate and appraise the effect of introducing political and technological measures to mitigate the impact of road traffic on society. In the second case, car-following models are mainly