Learning Motion Patterns of People for Compliant Robot Motion

By Maren Bennewitz, Wolfram Burgard, Grzegorz Cielniak, Sebastian Thrun. Int. J. Robotics Res., 2005

468 citations recorded by the source corpus.

Whenever people move through their environments they do not move randomly. Instead, they usually follow specific trajectories or motion patterns corresponding to their intentions. Knowledge about such patterns enables a mobile robot to robustly keep track of persons in its environment and to improve its behavior. In this paper we propose a technique for learning collections of trajectories that characterize typical motion patterns of persons. Data recorded with laser-range finders are clustered using the expectation maximization algorithm. Based on the result of the clustering process, we derive a hidden Markov model that is applied to estimate the current and future positions of persons based on sensory input. We also describe how to incorporate the probabilistic belief about the potential trajectories of persons into the path planning process of a mobile robot. We present several experiments carried out in different environments with a mobile robot equipped with a laser-range scanner and a camera system. The results demonstrate that our approach can reliably learn motion patterns of persons, can robustly estimate and predict positions of persons, and can be used to improve the navigation behavior of a mobile robot.

Subfields: Aerospace Engineering, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/2d22e4cf3009595b48d0439b448dc9266dfd33ea

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Aerospace Engineering

In Computer Vision and Pattern Recognition

Citations in the corpus

Social Attention: Modeling Attention in Human Crowds (2017, background): “to predict future trajectories as in [3], [19], [27].”
Crowd-Robot Interaction: Crowd-Aware Robot Navigation With Attention-Based Deep Reinforcement Learning (2018, background): “Earlier works separate prediction and planning in two steps, attempting to identify a safe path after forecasting the future trajectories of the others [16], [17].”
Unfreezing the robot: Navigation in dense, interacting crowds (2010, background): “Current approaches, such as [23], [2], [10], [14] develop human motion models to augment the prediction, in the hope that good prediction will lead to good navigation.”
Human motion trajectory prediction: a survey (2019, methodology): “…and possibly velocity), e.g. (Ferrer and Sanfeliu 2014; Elfring et al. 2014; Pellegrini et al. 2009; Kitani et al. 2012; Karasev et al. 2016; Ziebart et al. 2009; Kooij et al. 2018; Trautman and Krause 2010; Kuderer et al. 2012; Bennewitz et al. 2005; Kucner et al. 2017; Bera et al. 2016) 1.2.”
Intention-aware online POMDP planning for autonomous driving in a crowd (2015, unknown): “Our model is currently very basic, but we plan to build richer and more accurate probabilistic dynamic models through learning [2], [22] in the future.”
Intentional motion on-line learning and prediction (2008, methodology): “Finally, it is worth mentioning again [3] in this context.”
A Survey of Methods for Safe Human-Robot Interaction (2017, methodology): “An earlier approach by Bennewitz et al. (2005) uses no explicit motion models; instead, learning is performed automatically by clustering human motions via the expectation maximization (EM) algorithm and modeling these clustered motions using hidden Markov models (HMMs).”
Incremental Learning, Clustering and Hierarchy Formation of Whole Body Motion Patterns using Adaptive Hidden Markov Chains (2008, background): “…including skill transfer (Yang et al. 1997 Dillmann et al. 1999), robot-assisted surgery (Kragic et al. 2005 Ekvall et al. 2006), sign language and gesture modeling (Startner and Pentland 1995 Bernardin et al. 2005 Iba et al. 2005), and motion prediction (Bennewitz et al. 2005 Ho et al. 2005).”
Core Challenges of Social Robot Navigation: A Survey (2021, background): “Similarly, Bennewitz et al. [15] extracted patterns of human motion in a crowded environment and derive a hidden Markov model to perform online human motion prediction.”
People tracking with human motion predictions from social forces (2010, background): “This is in contrast to previous work that uses combinations of methods, such as goal learning and planning [5], Voronoi graph extraction from a map [6] or EM clustering and HMMs [7].”
Robot navigation in dense human crowds: Statistical models and experimental studies of human–robot cooperation (2015, background): “Indeed, if the amount of uncertainty was the primary reason for this suboptimal navigation, then using more precise individual dynamics models would prevent the FRP.”
Human-aware robot navigation: A survey (2013, background): “Naturalness is the similarity between robots and humans in low-level behavior patterns.”
Extracting Places and Activities from GPS Traces Using Hierarchical Conditional Random Fields (2007, background): “…the problem of determining a person’s signiﬁcant places by assuming that a geographic loca-119 tion is signiﬁcant if and only if the user spends at least θ minutes there, for some ﬁxed threshold θ (Ashbrook and Starner 2003; Liao et al. 2004, 2005c; Gogate et al. 2005; Bennewitz et al. 2005).”