Robotic Exploration for Learning Human Motion Patterns

Abstract

Understanding how people are likely to move is key to efficient and safe robot navigation in human environments. However, mobile robots can only observe a fraction of the environment at a time, while the activity patterns of people may also change at different times. This paper introduces a new methodology for mobile robot exploration to maximise the knowledge of human activity patterns by deciding where and when to collect observations. We introduce an exploration policy driven by the entropy levels in a spatio-temporal map of pedestrian flows, and compare multiple spatio-temporal exploration strategies including both informed and uninformed approaches. The evaluation is performed by simulating mobile robot exploration using real sensory data from three long-term pedestrian datasets. The results show that for certain scenarios the models built with proposed exploration system can better predict the flow patterns than uninformed strategies, allowing the robot to move in a more socially compliant way, and that the exploration ratio is a key factor when it comes to the model prediction accuracy.

Survey of maps of dynamics for mobile robots

“…In more recent years, another work tackling the problem of spatio-temporal exploration was published by Molina et al (2021). In this paper, the authors explore the problem of gathering observations for time-dependent flow map STeFMap (Molina et al, 2018).…”

Non-Parametric Modeling of Spatio-Temporal Human Activity Based on Mobile Robot Observations

“…where larger values indicate less accurate prediction compared with the test data. These two metrics can be regarded as the standard metrics when comparing human activity or flow models [3], [6], [16], [27]. However, these criteria have limited expressiveness in terms of the usefulness of the model e.g.…”

“…χ 2 -distance is given as multipliers of 10 4 for brevity of notation. 27. 2.34 32.03 2.57 35.75 2.77 38.89 1.79 456.92 1.94 460.32 1.84 477.26 2.0 481.37 Fr-AAM 2.6 33.72 2.38 30.76 2.61 33.89 2.67 34.89 1.36 345.00 1.31 438.58 1.32 420.39 1.28 429.97 GP-Hom 2.58 31.35 2.34 33.49 2.56 33.60 2.32 34.74 1.57 649.01 1.5 639.74 1.41 567.46 1.37 576.19 CoPA-Map 2.61 50.52 2.13 24.91 2.33 25.28 2.19 25.20 1.44 606.58 0.82 172.07 0.8 164.56 0.69 127.15…”

Non-Parametric Modeling of Spatio-Temporal Human Activity Based on Mobile Robot Observations

“…Office Dataset [27]: The second dataset contains tracks of people based on measurements by a single stationary 3D-Lidar in an office environment of the University of Lincoln, England, covering an area of ca. 85 m 2 with averagely about 300 entries per square meter and day.…”

“…χ 2 -distance is given as multipliers of 10 4 for brevity of notation. 27. 2.34 32.03 2.57 35.75 2.77 38.89 1.79 456.92 1.94 460.32 1.84 477.26 2.0 481.37 Fr-AAM 2.6 33.72 2.38 30.76 2.61 33.89 2.67 34.89 1.36 345.00 1.31 438.58 1.32 420.39 1.28 429.97 GP-Hom 2.58 31.35 2.34 33.49 2.56 33.60 2.32 34.74 1.57 649.01 1.5 639.74 1.41 567.46 1.37 576.19 CoPA-Map 2.61 50.52 2.13 24.91 2.33 25.28 2.19 25.20 1.44 606.58 0.82 172.07 0.8 164.56 0.69 127.15Copyright (c) 2022 IEEE.…”