Abstract
With the rise of autonomous public transportation, passenger safety in autonomous buses is paramount. This paper introduces a novel Multimodal Long Short-Term Memory (LSTM) network-based fall detection system, to enhance passenger safety by accurately detecting falls, thereby assisting remote supervisors in efficiently monitoring multiple vehicles. It comprises two main processes: feature extraction and fall discrimination. Feature extraction utilizes YOLO (You Only Look Once) v7 for real-time pose estimation, combined with the SORT algorithm for tracking individuals across video frames. Fall discrimination leverages sequential data processing with LSTM networks in the proposed Multimodal approach, which employs a fusion of pose estimation and Inertial Measurement Unit (IMU) sensor data. Evaluations conducted in various fall and non-fall scenarios within a realistic bus setting yielded high recall and F2 scores. Specifically, the model attained 98% recall in single person fall scenarios and 95% in more complex multi-person fall scenarios, significantly surpassing traditional single-modality approaches such as Multilayer Perceptron (MLP) and simple LSTM. The paper also investigates a decision-level fusion approach, balancing predictive accuracy by optimizing the late integration of separate Pose and IMU models, despite its higher computational cost. The impact of varying frame rates on model performance was also explored, addressing practical implications for real-world implementations. The robustness of the model was affirmed even at reduced processing frame rates, ensuring effective real-time processing capabilities crucial for ensuring timely responses in emergency situations. The development of the fall detection system promises a safer and more efficient future for public transportation, especially benefiting elderly individuals.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Alonso-González, M.J., Hoogendoorn-Lanser, S., van Oort, N., Cats, O., Hoogendoorn, S.: Drivers and barriers in adopting mobility as a service (MaaS) – a latent class cluster analysis of attitudes. Transp. Res. Part a Policy Pract. 132, 378–401 (2020)
Zhang, X., Ji, J., Wang, L., He, Z., Liu, S.: Image-based fall detection in bus compartment scene. IET Image Process. 17, 118–119 (2023)
Wang, C.-Y., Bochkovskiy, A., Liao, H.-Y.M.: YOLOv7: Trainable Bag-of-Freebies Sets New State-of-the-Art for Real-Time Object Detectors. in (2023). https://doi.org/10.1109/cvpr52729.2023.00721
Hochreiter, S., Schmidhuber, J.: Long Short-Term Memory. Neural Comput. 9, 1735–1780 (1997)
GarcÃa, E., et al.: Towards effective detection of elderly falls with CNN-LSTM neural networks. Neurocomputing 500, 231–240 (2022)
Wu, J., Wang, J., Zhan, A., Wu, C.: Fall detection with CNN-casual LSTM network. Information (Switzerland) 12, 403 (2021)
Salimi, M., Machado, J.J.M., Tavares, J.M.R.S.: Using deep neural networks for human fall detection based on pose estimation. Sensors 22, 4544 (2022)
Dakare, A.A., Wu, Y., Hashimoto, N., Kumagai, T., Miura, T.: Fall detection inside an autonomous driving bus: - examination of image processing algorithms. In: Digest of Technical Papers - IEEE International Conference on Consumer Electronics, vol. 2023, January 2023
Dakare, A.A., Wu, Y., Kumagai, T., Miura, T., Hashimoto, N.: Advancing fall detection in an autonomous bus: examination of LSTM technique. In: Proceedings of the International Conference on Mechanical and Robotics Engineering (ICMRE 2024). IEEE (2024, to be published)
Bewley, A., Ge, Z., Ott, L., Ramos, F., Upcroft, B.: Simple online and realtime tracking. In: Proceedings - International Conference on Image Processing, ICIP 2016-August, pp. 3464–3468 (2016)
Kingma, D.P., Ba, J.L.: Adam: a method for stochastic optimization. In: 3rd International Conference on Learning Representations, ICLR 2015 - Conference Track Proceedings (2015)
Acknowledgments
This research was supported by the Ministry of Economy, Trade and Industry, Japan and I would like to express my sincere gratitude to them for their generous support in facilitating this research.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Ethics declarations
Disclosure of Interests
The authors have no competing interests to declare that are relevant to the content of this article.
Rights and permissions
Copyright information
© 2024 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this paper
Cite this paper
Dakare, A.A., Wu, Y., Kumagai, T., Miura, T., Hashimoto, N. (2024). Enhancing Passenger Safety in an Autonomous Bus: A Multimodal Fall Detection Approach for Effective Remote Monitoring. In: Stephanidis, C., Antona, M., Ntoa, S., Salvendy, G. (eds) HCI International 2024 Posters. HCII 2024. Communications in Computer and Information Science, vol 2118. Springer, Cham. https://doi.org/10.1007/978-3-031-61963-2_17
Download citation
DOI: https://doi.org/10.1007/978-3-031-61963-2_17
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-61962-5
Online ISBN: 978-3-031-61963-2
eBook Packages: Computer ScienceComputer Science (R0)