Probabilistic Mapping of Spatial Motion Patterns for Mobile Robots

Home / Publications / 2020 / Probabilistic Mapping of Spatial Motion Patterns for Mobile Robots

Tomasz Piotr Kucner, Achim J. Lilienthal, Martin Magnusson, Luigi Palmieri, Chittaranjan Srinivas Swaminathan
Probabilistic Mapping of Spatial Motion Patterns for Mobile Robots
Springer, Cognitive Systems Monographs (40)

 

Preface

This book focuses on the topic of building spatial models that show patterns of dynamics in the environment and using these models for solving the problem of mobile robot motion planning in populated environments. In this book, we especially explore probabilistic methods for modelling the flow of discrete objects (i.e. people) and continuous media (i.e. air), based on sparse and noisy data.

Motion and change are inherent parts of the world. Yet robotic research has often assumed for reasons of practicability that robots operate in static environments. In cases where the presence of dynamic changes in the environment of a robot cannot be ignored, the dynamics is predominantly addressed in a local context, for instance, by motion prediction algorithms that anticipate the future position of objects perceived by the robot. However, such approaches neglect global information about patterns of dynamics, which we as humans use routinely.

Maps of the environment provide information allowing robots to reason about parts of the environment beyond the reach of their sensors. Yet, robotic mapping has traditionally focused on the static features of the environment while paying little attention to the task of building spatial models of dynamics.

In this book, we provide an introduction to the problem of building maps of dynamics (MoDs), followed by an overview of the existing body of work. We present two methods for building flow maps in detail. The first approach models dynamics as conditional probability of occupancy shift over the grid map. The second approach model dynamics as field of Gaussian mixture models describing the local distribution of velocities. We also demonstrate how maps of dynamics can be used in motion planning. Finally, we present possible directions for further research related to maps of dynamics.

https://link.springer.com/book/10.1007%2F978-3-030-41808-3

@Book{kucner-2020-probabilistic,
 author = {Tomasz Piotr Kucner and Achim J. Lilienthal and Martin Magnusson and Luigi Palmieri and Chittaranjan Srinivas Swaminathan},
 publisher = {Springer International Publishing},
 title = {Probabilistic Mapping of Spatial Motion Patterns for Mobile Robots},
 year = {2020},
 series = {Cognitive Systems Monographs},
 volume = {40},
 doi = {10.1007/978-3-030-41808-3},
}