Autonomous Driving with Interpretable Goal Recognition and Monte Carlo Tree Search

Cillian Brewitt; Stefano V Albrecht; John Wilhelm; Balint Gyevnar; Francisco Eiras; Mihai Dobre; Ram Ramamoorthy

Autonomous Driving with Interpretable Goal Recognition and Monte Carlo Tree Search

Cillian Brewitt, Stefano V Albrecht, John Wilhelm, Balint Gyevnar, Francisco Eiras, Mihai Dobre, Ram Ramamoorthy

Research output: Contribution to conference › Paper › peer-review

Abstract / Description of output

Abstract—The ability to predict the intentions and driving trajectories of other vehicles is a key problem for autonomous driving. We propose an integrated planning and prediction system which leverages the computational benefit of using a finite space of maneuvers, and extend the approach to planning and prediction of sequences of maneuvers via rational inverse planning to recognise the goals of other vehicles. Goal recognition informs a Monte Carlo Tree Search (MCTS) algorithm to plan optimal maneuvers for the ego vehicle. Our system constructs plans which are explainable by means of rationality. Evaluation in simulations of four urban driving scenarios demonstrate the system’s ability to robustly recognise the goals of other vehicles while generating near-optimal plans.

Original language	English
Number of pages	6
Publication status	E-pub ahead of print - 13 Jul 2020
Event	Interaction and Decision-Making in Autonomous-Driving: A Virtual Workshop at RSS 2020 - Duration: 13 Jul 2020 → 13 Jul 2020 https://sites.google.com/view/ida2020

Workshop

Workshop	Interaction and Decision-Making in Autonomous-Driving
Abbreviated title	IDA 2020
Period	13/07/20 → 13/07/20
Internet address	https://sites.google.com/view/ida2020

Access to Document

Autonomous driving with interpretable_BREWITT_DOA02072020_AFVAccepted author manuscript, 506 KBLicence: All Rights Reserved

https://drive.google.com/file/d/1dIWMbcAN35v8mCCcccWDbbBo7pSbKGuW/viewLicence: All Rights Reserved

Fast, Accurate, and Safe Behaviour Prediction in Autonomous Vehicles
Albrecht, S.
UK-based charities
1/01/19 → 31/12/22
Project: Research

Cite this

@conference{dad1c1afdda442568eea5b25b055a4e3,

title = "Autonomous Driving with Interpretable Goal Recognition and Monte Carlo Tree Search",

abstract = "Abstract—The ability to predict the intentions and driving trajectories of other vehicles is a key problem for autonomous driving. We propose an integrated planning and prediction system which leverages the computational benefit of using a finite space of maneuvers, and extend the approach to planning and prediction of sequences of maneuvers via rational inverse planning to recognise the goals of other vehicles. Goal recognition informs a Monte Carlo Tree Search (MCTS) algorithm to plan optimal maneuvers for the ego vehicle. Our system constructs plans which are explainable by means of rationality. Evaluation in simulations of four urban driving scenarios demonstrate the system{\textquoteright}s ability to robustly recognise the goals of other vehicles while generating near-optimal plans.",

author = "Cillian Brewitt and Albrecht, {Stefano V} and John Wilhelm and Balint Gyevnar and Francisco Eiras and Mihai Dobre and Ram Ramamoorthy",

year = "2020",

month = jul,

day = "13",

language = "English",

note = "Interaction and Decision-Making in Autonomous-Driving : A Virtual Workshop at RSS 2020, IDA 2020 ; Conference date: 13-07-2020 Through 13-07-2020",

url = "https://sites.google.com/view/ida2020",

}

TY - CONF

T1 - Autonomous Driving with Interpretable Goal Recognition and Monte Carlo Tree Search

AU - Brewitt, Cillian

AU - Albrecht, Stefano V

AU - Wilhelm, John

AU - Gyevnar, Balint

AU - Eiras, Francisco

AU - Dobre, Mihai

AU - Ramamoorthy, Ram

PY - 2020/7/13

Y1 - 2020/7/13

N2 - Abstract—The ability to predict the intentions and driving trajectories of other vehicles is a key problem for autonomous driving. We propose an integrated planning and prediction system which leverages the computational benefit of using a finite space of maneuvers, and extend the approach to planning and prediction of sequences of maneuvers via rational inverse planning to recognise the goals of other vehicles. Goal recognition informs a Monte Carlo Tree Search (MCTS) algorithm to plan optimal maneuvers for the ego vehicle. Our system constructs plans which are explainable by means of rationality. Evaluation in simulations of four urban driving scenarios demonstrate the system’s ability to robustly recognise the goals of other vehicles while generating near-optimal plans.

AB - Abstract—The ability to predict the intentions and driving trajectories of other vehicles is a key problem for autonomous driving. We propose an integrated planning and prediction system which leverages the computational benefit of using a finite space of maneuvers, and extend the approach to planning and prediction of sequences of maneuvers via rational inverse planning to recognise the goals of other vehicles. Goal recognition informs a Monte Carlo Tree Search (MCTS) algorithm to plan optimal maneuvers for the ego vehicle. Our system constructs plans which are explainable by means of rationality. Evaluation in simulations of four urban driving scenarios demonstrate the system’s ability to robustly recognise the goals of other vehicles while generating near-optimal plans.

M3 - Paper

T2 - Interaction and Decision-Making in Autonomous-Driving

Y2 - 13 July 2020 through 13 July 2020

ER -

Autonomous Driving with Interpretable Goal Recognition and Monte Carlo Tree Search

Abstract / Description of output

Workshop

Access to Document

Fingerprint

Projects

Fast, Accurate, and Safe Behaviour Prediction in Autonomous Vehicles

Cite this