Generating accurate and efficient predictions for the motion of the humans present in the scene is key to the development of effective motion planning algorithms for robots moving in promiscuous areas, where wrong planning decisions could generate safety hazard or simply make the presence of the robot 'socially' unacceptable. Our approach to predict human motion is based on a neural network of a peculiar kind. Contrary to conventional deep neural networks, our network embeds in its structure the popular Social Force Model, a dynamic equation describing the motion in physical terms. This choice allows us to concentrate the learning phase in the aspects which are really unknown (i.e., the model's parameters) and to keep the structure of the network simple and manageable. As a result, we are able to obtain a good prediction accuracy even by using a small and synthetically generated training set. Importantly, the prediction accuracy remains acceptable even when the network is applied in scenarios radically different from those for which it was trained. Finally, the choices of the network are 'explainable', as they can be interpreted in physical terms. Comparative and experimental results prove the effectiveness of the proposed approach.

Efficient Prediction of Human Motion for Real-Time Robotics Applications with Physics-Inspired Neural Networks / Antonucci, Alessandro; Rosati Papini, Gastone Pietro; Bevilacqua, Paolo; Palopoli, Luigi; Fontanelli, Daniele. - In: IEEE ACCESS. - ISSN 2169-3536. - STAMPA. - 10:(2022), pp. 144-157. [10.1109/ACCESS.2021.3138614]

Efficient Prediction of Human Motion for Real-Time Robotics Applications with Physics-Inspired Neural Networks

Antonucci, Alessandro;Rosati Papini, Gastone Pietro;Bevilacqua, Paolo;Palopoli, Luigi;Fontanelli, Daniele
2022

Abstract

Generating accurate and efficient predictions for the motion of the humans present in the scene is key to the development of effective motion planning algorithms for robots moving in promiscuous areas, where wrong planning decisions could generate safety hazard or simply make the presence of the robot 'socially' unacceptable. Our approach to predict human motion is based on a neural network of a peculiar kind. Contrary to conventional deep neural networks, our network embeds in its structure the popular Social Force Model, a dynamic equation describing the motion in physical terms. This choice allows us to concentrate the learning phase in the aspects which are really unknown (i.e., the model's parameters) and to keep the structure of the network simple and manageable. As a result, we are able to obtain a good prediction accuracy even by using a small and synthetically generated training set. Importantly, the prediction accuracy remains acceptable even when the network is applied in scenarios radically different from those for which it was trained. Finally, the choices of the network are 'explainable', as they can be interpreted in physical terms. Comparative and experimental results prove the effectiveness of the proposed approach.
Antonucci, Alessandro; Rosati Papini, Gastone Pietro; Bevilacqua, Paolo; Palopoli, Luigi; Fontanelli, Daniele
Efficient Prediction of Human Motion for Real-Time Robotics Applications with Physics-Inspired Neural Networks / Antonucci, Alessandro; Rosati Papini, Gastone Pietro; Bevilacqua, Paolo; Palopoli, Luigi; Fontanelli, Daniele. - In: IEEE ACCESS. - ISSN 2169-3536. - STAMPA. - 10:(2022), pp. 144-157. [10.1109/ACCESS.2021.3138614]
File in questo prodotto:
File Dimensione Formato  
Efficient_Prediction_of_Human_Motion_for_Real-Time_Robotics_Applications_With_Physics-Inspired_Neural_Networks(1).pdf

accesso aperto

Tipologia: Versione editoriale (Publisher’s layout)
Licenza: Creative commons
Dimensione 2.24 MB
Formato Adobe PDF
2.24 MB Adobe PDF Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione

Utilizza questo identificativo per citare o creare un link a questo documento: http://hdl.handle.net/11572/328850
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 1
  • ???jsp.display-item.citation.isi??? 0
social impact