Modelling Groups of Humans: Towards Crowd Digital Twins

Modelling Groups of Humans: Towards Crowd Digital Twins

• 1.
  Modelling Groups ofHumans: Towards Crowd Digital Twins Roberto Casadei, Giovanni Delnevo, Roberto Girau, Silvia Mirri Department of Computer Science and Engineering (DISI) ALMA MATER STUDIORUM – Università of Bologna June 26, 2024 Workshop on Social Media Sensing (SMS) https://www.slideshare.net/RobertoCasadei/ R. Casadei Motivation Contribution Wrap-up References 1/12
• 2.
  Overview a crowd digitaltwin (Crowd-DT) is a digital twin of a physical crowd (group of people) physical environment physical crowd environmental sensors (e.g., cameras) stakeholders crowd digital twin data warehouse models prediction planning environment digital twin external services wearables / devices In the workshop paper, we address four main questions: ? what motivates (investigating on) a Crowd-DT concept? ? what defines a Crowd-DT, and what requirements/desiderata follow? ? what enablers and challenges to implementation? ? how applications can be supported by a working Crowd-DT system? R. Casadei Motivation Contribution Wrap-up References 2/12
• 3.
  Outline 1 Motivation 2 Contribution 3Wrap-up
• 4.
  Business motivation –why (investigating) Crowd-DTs? Business needs / Application domains ? What business needs or potential applications motivate the idea of addressing physical crowds in software? a classification of crowd-related services 1. services targetting crowds (participants) (cf. crowd mitigation, crowd emergency management) 2. services supported by crowds (cf. crowdsourcing) 3. services that should be crowd-aware (cf. navigating an environment while avoiding overcrowded areas) reference scenario: crowd management ∠ monitoring and controlling the members of the crowd and environmental elements ∠ pandemic prevention/mitigation, emergency management... R. Casadei Motivation Contribution Wrap-up References 3/12
• 5.
  Technical motivation –why (investigating) CDTs? Research trends ? What scientific background motivates the idea of addressing physical crowds in software with a digital twin-based solution? Foundations research has demonstrated the opportunity and feasibiliy of crowd modelling ∠ physics of crowd dynamics ∠ psychology of crowds research on complex/collective systems and collective intelligence ∠ considering collectives as first-class entities Related work human digital twins digital twins of groups/swarms R. Casadei Motivation Contribution Wrap-up References 4/12
• 6.
  Outline 1 Motivation 2 Contribution 3Wrap-up
• 7.
  Crowd-DT Definition a crowd digitaltwin (CDT)... ... is the digital twin associated to a physical crowd, namely an entire group of co-located humans (representing its physical twin), ... ... aimed at supporting visualisation, analysis, and control of the system... ... through the definition of a digital-physical link (possibly bidirectional) with: 1. the (members of the) crowd 2. the surrounding environment, and 3. the associated collective phenomena. physical environment physical crowd environmental sensors (e.g., cameras) stakeholders crowd digital twin data warehouse models prediction planning environment digital twin external services wearables / devices R. Casadei Motivation Contribution Wrap-up References 5/12
• 8.
  Crowd-DT Typical Requirements /Desiderata 1. multi-scale — from micro to macro [1] 2. context-awareness — by connection with an environment DT [2] 3. real-time support — whose softness/hardness depending on the app 4. predictive capabilities — to understand future and mitigate risks 5. operational robustness — by fault-tolerance, graceful degradation, and prioritisation of safety-critical services 6. security and privacy — cf. CIA triad 7. bidirectional synchronisation — by “actuation” on the crowds and the environment (e.g., [3]) [1] A. Corbetta and F. Toschi, “Physics of human crowds,” Annu. Rev. Condens. Matter Phys., no. 1, 2023 [2] M. Maimour, A. Ahmed, and E. Rondeau, “Survey on digital twins for natural environments: A communication network perspective,” Internet of Things, Apr. 2024 [3] C. Feliciani, S. Tanida, X. Jia, and K. Nishinari, “Influencing pedestrian route choice through environmental stimuli: A long-term ecological experiment,” J. Disaster Res., no. 2, Apr. 2024 R. Casadei Motivation Contribution Wrap-up References 6/12
• 9.
  Crowd-DT Enablers models of crowdbehaviour (from sub-micro to super-macro scale) [1] ∠ cf. molecular dynamics, statistical mechanics, and fluid dynamics sensor fusion to efficiently integrate data from multiple sources ∠ cf. sheaf theory to map local to global edge computing to support low-latency computational services crowd simulators [4] people/environmental sensors/actuators [3] and works on emergence steering [5] collective intelligence/computing [6] and macro-programming [7] [1] A. Corbetta and F. Toschi, “Physics of human crowds,” Annu. Rev. Condens. Matter Phys., no. 1, 2023 [4] S. Yang, T. Li, X. Gong, B. Peng, and J. Hu, “A review on crowd simulation and modeling,” Graphical Models, 2020 [3] C. Feliciani, S. Tanida, X. Jia, and K. Nishinari, “Influencing pedestrian route choice through environmental stimuli: A long-term ecological experiment,” J. Disaster Res., no. 2, Apr. 2024 [5] R. Casadei, D. Pianini, M. Viroli, and D. Weyns, “Digital twins, virtual devices, and augmentations for self-organising cyber-physical collectives,” Applied Sciences, no. 1, 2022 [6] R. Casadei, “Artificial collective intelligence engineering: A survey of concepts and perspectives,” Artif. Life, no. 4, 2023 [7] R. Casadei, “Macroprogramming: Concepts, state of the art, and opportunities of macroscopic behaviour modelling,” ACM Comput. Surv., no. 13s, 2023 R. Casadei Motivation Contribution Wrap-up References 7/12
• 10.
  Crowd-DT: use cases Small-scalecrowd management: Campus and Public Transport Smart campus emergency management during campus evacuation Smart mobility surveillance suggestions for efficient mobility (e.g., suggest routes to prevent congestion) R. Casadei Motivation Contribution Wrap-up References 8/12
• 11.
  Crowd-DT: use cases Large-scalecrowd management in mass events [8] crowd-aware navigation crowd dispersal suggestions crowd evacuation support [8] R. Casadei, G. Fortino, D. Pianini, W. Russo, C. Savaglio, and M. Viroli, “A development approach for collective opportunistic edge-of-things services,” Inf. Sci., 2019 R. Casadei Motivation Contribution Wrap-up References 9/12
• 12.
  Crowd-DT: challenges 1. Low-latencysynchronisation and prediction of crowd behaviour 2. Actuation on crowds 3. Interdisciplinary design 4. Adaptation to available and heterogeneous infrastructure (cf. edge-cloud continuum) 5. Heterogeneous crowds (cf. human-machine networks) R. Casadei Motivation Contribution Wrap-up References 10/12
• 13.
  Outline 1 Motivation 2 Contribution 3Wrap-up
• 14.
  Conclusion and FutureWork Conclusion: the notion of Crowd-DT appears to be both useful and feasible What: Crowd-DT as a promising solution for crowd management in smart environments Enabled by: research on crowd dynamics, sensor fusion, simulation, collective computing Challenges: related to real-time, actuation, interdisciplinary design, and heterogeneity Future work Reference architecture for Crowd-DT Simulation of Crowd-DT for the Smart Campus scenario Real-world prototype at the UNIBO campus R. Casadei Motivation Contribution Wrap-up References 11/12
• 15.
  References (1/1) [1] A.Corbetta and F. Toschi, “Physics of human crowds,” Annu. Rev. Condens. Matter Phys., vol. 14, no. 1, 311–333, 2023, ISSN: 1947-5462. DOI: 10.1146/annurev-conmatphys-031620-100450. [Online]. Available: http://dx.doi.org/10.1146/annurev-conmatphys-031620-100450. [2] M. Maimour, A. Ahmed, and E. Rondeau, “Survey on digital twins for natural environments: A communication network perspective,” Internet of Things, vol. 25, p. 101 070, Apr. 2024, ISSN: 2542-6605. DOI: 10.1016/j.iot.2024.101070. [Online]. Available: http://dx.doi.org/10.1016/j.iot.2024.101070. [3] C. Feliciani, S. Tanida, X. Jia, and K. Nishinari, “Influencing pedestrian route choice through environmental stimuli: A long-term ecological experiment,” J. Disaster Res., vol. 19, no. 2, 325–335, Apr. 2024, ISSN: 1881-2473. DOI: 10.20965/jdr.2024.p0325. [Online]. Available: http://dx.doi.org/10.20965/jdr.2024.p0325. [4] S. Yang, T. Li, X. Gong, B. Peng, and J. Hu, “A review on crowd simulation and modeling,” Graphical Models, vol. 111, p. 101 081, 2020, ISSN: 1524-0703. DOI: 10.1016/j.gmod.2020.101081. [Online]. Available: http://dx.doi.org/10.1016/j.gmod.2020.101081. [5] R. Casadei, D. Pianini, M. Viroli, and D. Weyns, “Digital twins, virtual devices, and augmentations for self-organising cyber-physical collectives,” Applied Sciences, vol. 12, no. 1, 2022, ISSN: 2076-3417. DOI: 10.3390/app12010349. [Online]. Available: https://doi.org/10.3390/app12010349. [6] R. Casadei, “Artificial collective intelligence engineering: A survey of concepts and perspectives,” Artif. Life, vol. 29, no. 4, pp. 433–467, 2023. DOI: 10.1162/ARTL_A_00408. [Online]. Available: https://doi.org/10.1162/artl_a_00408. [7] R. Casadei, “Macroprogramming: Concepts, state of the art, and opportunities of macroscopic behaviour modelling,” ACM Comput. Surv., vol. 55, no. 13s, 275:1–275:37, 2023. DOI: 10.1145/3579353. [Online]. Available: https://doi.org/10.1145/3579353. [8] R. Casadei, G. Fortino, D. Pianini, W. Russo, C. Savaglio, and M. Viroli, “A development approach for collective opportunistic edge-of-things services,” Inf. Sci., vol. 498, pp. 154–169, 2019. DOI: 10.1016/J.INS.2019.05.058. [Online]. Available: https://doi.org/10.1016/j.ins.2019.05.058. R. Casadei Motivation Contribution Wrap-up References 12/12