Experiments toward an human-like Adaptive Cruise Control

Abstract

In this work some experiments are made in order to assess the feasibility of a human-like ACC (Adaptive Cruise Control) system. The proposed system is able to understand driver’s attitudes and driving-styles by means of a selfcalibration process that can be (re)initialized on request. Three different speed-control logics have been tested: one tries to learn from actual drivers’ behaviors by using an Artificial eural etwork (A) approach, the second is based on the calibration of a linear function aimed to be mimic of the driver response to stimuli, the third is based on the calibration of a polynomial function instead of a linear one. A microscopic traffic model, accurately calibrated and validated for different aims and in a previous work, has been adapted and used in order to generate a long car-following trajectory on which the speed control logics have been calibrated and compared. This has allowed for a sufficient amount of accurate laboratory data at a relatively low cost. Comparison of the tested speed-control logics show that a fully adaptive human-like ACC system is feasible and worth further more costly developments

Driving behaviour models enabling the simulation of Advanced Driving Assistance Systems: revisiting the Action Point paradigm

“…In stimulus-based models the acceleration of the follower is argued to be determined by the reaction to the relative velocity and spacing to the vehicle in front. Perhaps the best known examples of these are the series of relationships formulated in the late 1950s as part of the so-called General Motors (GM) model (Chandler et al, 1958;Gazis et al, 1959;Herman et al, 1959), which has continued in use to the present day (e.g., Ozaki, 1993;Bifulco et al, 2008). The safety-distance approach was developed by arguing that a follower retains a distance and a speed that allows for safe braking if the leader slows abruptly.…”

“…Arguably, the first were those used in Collision Warning which focussed on the use of Time To Collision (TTC, the ratio of rear to front separation to relative speed, see Abe and Richardson, 2006 for example). Such approaches have more recently been used in ACC design, where studies range from empirical investigations (e.g., Wang et al, 2011, andBoyle, 2013), to modelling exercises (e.g., Bifulco et al, 2008, who proposed implementation of a following model using a machine learning technique), through to large-scale experimental FOT (Field Operational Test) studies, such as those reported in Viti et al (2008). It can be seen, therefore, that studies and models of driver behaviour (in particular during car following) are crucial to making advancements in ADAS (Simonelli et al, 2009), and this is a rich field undergoing rapid development (e.g., Colombaroni and Fusco, 2013, who propose an Artificial Neural Network approach for traffic-simulation oriented car-following models).…”

Longitudinal control behaviour: Analysis and modelling based on experimental surveys in Italy and the UK

“…In stimulus-based models the acceleration of the follower is determined by the reaction to 32 the relative velocity and spacing from the leading vehicle. Such models were made popular by the paradigms 33 formulated in the late 1950s as part of the so-called General Motors (GM) model (Gazis et al, 1961), which has 34 continued in use to the present day (Bifulco et al, 2008). The safety distance approach was developed by arguing that a 35 follower maintains a distance and a speed that allow safe braking if the leader slows abruptly.…”

Heterogeneity of Driving Behaviors in Different Car-Following Conditions

“…It is directly derived from a time-discrete stimulus-response approach and it is specifically developed for ADAS-oriented solutions. Once properly calibrated, it has been shown to be mimic of real car-following behaviors, as it is aimed at enabling an human-like control logic for adaptive cruise control (ACC) systems (Bifulco et al, 2008). The time-discrete linear model uses a polynomial formula to relate the instantaneous speeds of the leader and of the follower (and their relative spacing) with the target spacing the follower applies for the next simulation time-step.…”