Visually-guided motion planning for autonomous driving from interactive demonstrations

نویسندگان

چکیده

The successful integration of autonomous robots in real-world environments strongly depends on their ability to reason from context and take socially acceptable actions. Current navigation systems mainly rely geometric information hard-coded rules induce safe compliant behaviors. Yet, unstructured urban scenarios these approaches can become costly suboptimal. In this paper, we introduce a motion planning framework consisting two components: data-driven policy that uses visual inputs human feedback generate driving behaviors (encoded by high-level decision variables), local trajectory optimization method executes (ensuring safety). particular, employ Interactive Imitation Learning jointly train the with planner, Model Predictive Controller (MPC), which results human-like Our approach is validated realistic simulated scenarios. Qualitative show similarity learned driving. Furthermore, performance substantially improved terms safety, i.e., number collisions, as compared prior frameworks, data-efficiency learning-based broadening operational domain MPC more • behavior optimization-based planners. Trajectory planners are safer socially-compliant when they resemble behavior. allows learn few hours teaching.

برای دانلود رایگان متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Safe Motion Planning for Autonomous Driving

Self-driving cars have the potential to revolutionize transportation by making it cheaper, safer, and more efficient. In this thesis we describe a novel motion planning system, which translates high-level navigation goals into low-level actions for controlling a vehicle. Specifically, the motion planning system is responsible for choosing at each time step an appropriate velocity and steering a...

متن کامل

A Lightweight Simulator for Autonomous Driving Motion Planning Development

A good simulation environment will facilitate motion planning algorithm development for urban autonomous driving. The first requirement of such a simulator is to be able to replicate a complex urban environment, including road network, curb, general objects, etc. The second requirement is to simulate a realistic host vehicle, which includes perception, control and vehicle dynamics, to recreate ...

متن کامل

Interactive Autonomous Driving through Adaptation from Participation

We present an intelligent driving software system that is capable of adapting to dynamic task conditions and adjusting to the steering preferences of a human passenger. This interactive autonomous system is a realization of a natural human-robot interaction paradigm known as Adaptation from Participation (AfP). AfP exploits intermittent input from the human passenger in order to adapt the vehic...

متن کامل

Humanoid Locomotion Planning for Visually Guided Tasks

*Centro de Investigaci on en Matem aticas, CIMAT, Jalisco s/n. Col. Valenciana, 36240 Guanajuato, Guanajuato, M exico Departamento de Matem aticas, Universidad de Guanajuato, Jalisco s/n. Col. Valenciana, 36240 Guanajuato, Guanajuato, M exico Robotics and Advanced Manufacturing Group, Centro de Investigaci on y de Estudios Avanzados del IPN, CINVESTAV, Saltillo, Coah, M exico §LAAS-CNRS, Labora...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

ژورنال

عنوان ژورنال: Engineering Applications of Artificial Intelligence

سال: 2022

ISSN: ['1873-6769', '0952-1976']

DOI: https://doi.org/10.1016/j.engappai.2022.105277