Within this work a novel method was introduced to generate a trajectory for an autonomous race vehicle for both online avoidance and offline applications. The presented algorithm calculates high speed race optimal trajectories which enable an autonomous vehicle to drive at the limits of its friction capabilities. A great advantage of this method is its flexibility. A race line for the whole track can be computed offline, or trajectories can be calculated online for a smaller look ahead distance, which is crucial for obstacle avoidance. For the connection of the points in the state lattice, quintic polynomials are used to describe the offset to the reference line along the track distance. Each lattice point was then constrained to have a heading and curvature aligned with the reference line. A drawback of this solution is that the sign of the curvature derivative changes frequently, which results in suboptimal steering behavior. Including curvature and heading discretization at every point in the state lattice could solve this problem, but would result in a much higher dimensional search space. Another solution would be to perform a post optimization of the race line after the graph search. This method is proposed by Xu et al., to smooth the path and the velocity profile. With the velocity profile calculation proposed in this paper, an optimization of the path would be sufficient because a smooth path would automatically lead to a smooth velocity profile. For integration into the research vehicle, the first step will be the combination of the velocity profile calculation, described in this work, the offline computed race line, and the optimal control method for online obstacle avoidance. The next steps towards a full online trajectory planner will be the implementation and the optimization of the code for runtime.


    Access

    Access via TIB

    Check availability in my library


    Export, share and cite



    Title :

    Trajectory planning for high dynamic automated driving




    Publication date :

    2014


    Size :

    16 Seiten, Bilder, 10 Quellen


    Type of media :

    Conference paper


    Type of material :

    Storage medium


    Language :

    English




    Trajectory planning for high dynamic automated driving

    Keller,T. / Wimmershoff,M. / Hardy,J. et al. | Automotive engineering | 2014


    Model Predictive Trajectory Planning for Automated Driving

    Yi, Boliang / Bender, Philipp / Bonarens, Frank et al. | IEEE | 2019


    Dynamic Trajectory Planning Algorithm for Automated Driving for Unmanned Ground Vehicles

    Thakur, Shivani / Sharma, Virender | BASE | 2017

    Free access

    Dynamic Trajectory Planning for Vehicle Autonomous Driving

    Zhang, Sumin / Deng, Weiwen / Zhao, Qingrong et al. | IEEE | 2013