Page 24 - ASME DSCC 2015 Program

Technical Program

a Sequential Two-Step algorithm for fast generation of Vehicle racing
ConTRIBuTED SESSIon
1-28-1 WA6 Vehicle Dynamics Control 1 Trajectories
Emerson Burkhart B 10:00am–12:00pm Contributed regular paper. DSCC2015-9757
nitin Kapania, John Subosits, J Christian Gerdes, Stanford University,
Session Chair: Hui Zhang, Tsinghua University Stanford, CA, United States
Session Co-Chair: Yu Wang, Seagate Technology The problem of maneuvering a vehicle through a race course in minimum
time requires computation of both longitudinal (brake and throttle) and
evaluation of a Multi-Vehicle Merging Strategy under different Lateral
lateral (steering wheel) control inputs. Unfortunately, solving the resulting
Maneuvers in the Presence of Sudden Braking
nonlinear optimal control problem is typically computationally expensive and
Contributed regular paper. DSCC2015-9985
infeasible for real-time trajectory planning. This paper presents an iterative
Mohammad Goli, Azim Eskandarian, Virginia Tech, Blacksburg, VA, United algorithm that divides the path generation task into two sequential sub-
States
problems that are significantly easier to solve. Given an initial path through
The ability for multiple vehicles to merge into an ongoing platoon is an the race track, the algorithm runs a forward-backward integration scheme
important task in the field of intelligent transportation system (ITS). For such to determine the minimum-time longitudinal speed profile, subject to tire
task, first a strategy is required to include a set of rules and actions which friction constraints. With this speed profile fixed, the algorithm updates the
allows the incoming vehicles to join the platoon, and second; a safe lateral vehicle’s path by solving a convex optimization problem that minimizes the
trajectory generator which always ensures the safety of lateral maneuver in resulting path curvature while staying within track boundaries and obeying
terms of lateral acceleration. In this paper we evaluated the multi-vehicle affine, time-varying vehicle dynamics constraints. This two-step process is
merging strategy using different lateral trajectory functions under sudden repeated iteratively until the predicted lap time no longer improves. While
breaking. Simulation results show that for merging task, a switching function providing no guarantees of convergence or a globally optimal solution, the
is necessary to switch from an adaptive lateral function to a constant lateral approach performs well when tested on the Thunderhill Raceway course in
trajectory function in case of sudden decelerations Willows, CA. The lap time reaches a minimum value after only three itera-
tions, with each iteration over the full 5 km race course requiring only thirty
an adaptive Nonlinear differentiable friction Modeling for Tire-road
seconds of computation time on a laptop computer. The resulting vehicle
friction Estimation
path and speed profile match very well with a nonlinear gradient descent
Contributed regular paper. DSCC2015-9755
solution and a path driven by a professional racecar driver,indicating that the
Zhijun fu, xiaobin Ning, Zhejiang University of Technology, Hangzhou, proposed method is a viable option for online trajectory planning in the near
China, Subhash rakheja, Wen-fang xie, Concordia University, Montreal, future.
QC, Canada, Weidong xie, Zhejiang University of Technology, Zhejiang,
China Automated Robust Path following Control based on Calculation of
lateral Deviation and Yaw Angle Error
In this paper, a differentiable friction model is proposed to estimate the
Contributed regular paper. DSCC2015-9856
longitudinal tire-road friction force of vehicle systems. A novel adaptive non-
linear observer-based parameter estimation scheme has been developed Mumin Tolga Emirler, The Ohio State University Center for
to estimate the parameters of friction model, which requires the signals from Automotive Research, Columbus, OH, United States, Haoan Wang,
Bilin Aksun-Guvenc, levent Guvenc, The Ohio State University,
the existing sensors signals such as wheel rotational speed and vehicle
Columbus, OH, United States
speed. Different from conventional gradient and recursive least square (RLS)
methods, the filtered regression parameter estimation error is introduced in Automated driving vehicles are expected to be ready for series production
the novel adaptive laws, which can guarantee the observer error conver- by 2020. An important component of automated driving technologies is
gence to zero and the estimated parameter also convergence to their real controlled path following under longitudinal speed control. In this paper,
value. The Lyapunov method is used to prove the stability of the proposed a robust path following controller design based on lateral deviation and
methods. The robustness of the developing method against bounded distur- yaw angle error determination of the vehicle is proposed. The constrained
bances is also proved. Simulation results illustrate that the proposed method least square method is used for obtaining continuity and smoothness of the
can realize relatively accurate estimation of the friction with variations in segment boundaries of the digital trajectory map to be followed. The lateral
speed and road gradient. deviation and yaw angle error are calculated by comparing the generat-
ed digital map trajectory and the vehicle position. The parameter space
approach is used in the design stage of the controller considering D-stability
requirements. The solution regions of the controller are plotted in three
dimensional parameter space. The designed controller is tested with simula-
tions on a path chosen from the Ohio State University campus.

24