Periodic Reporting for period 1 - BITS (Bayesian Uncertainty Quantification of Intelligent vehicles)
Período documentado: 2020-06-01 hasta 2022-05-31
The project transfers Bayesian learning tools in the area of Intelligent Transportation Systems (ITS) and develops novel methodologies for quantifying the uncertainty in intelligent vehicles (IVs).
The main objectives of the project are:
1) Formulate the analysis of IVs in a probabilistic framework and develop methodologies for the representation of uncertainty.
2) Extend the framework for scenarios with several interacting IVs.
3) Develop algorithms for efficient decision-making and optimization, for individual IVs and a network of IVs.
4) Experimentation and validation of the developed methodologies to the physical testbed, a small-scale road network.
5) Support the ER to become a leading expert in the area of ITS and to transfer her acquired expertise to research institutions, companies and the public.
The project dealt at a theoretical level with uncertainty quantification and estimation of traffic variables. The proposed methods have shed light on the cutting-edge research areas of Bayesian inference, fault diagnosis and ITS. The main conclusions of the action are: (i) probabilistic traffic state estimation using data from IVs can reduce uncertainty up to 4 times compared to non-probabilistic methods, (ii) probabilistic traffic state estimation provides high-quality results even for low penetration rates of IVs, (iii) OD matrix estimation using measurements collected from IVs results in significantly better performance compared to measurements collected from fixed location sensors.
The main objectives of the project are:
1) Formulate the analysis of IVs in a probabilistic framework and develop methodologies for the representation of uncertainty.
2) Extend the framework for scenarios with several interacting IVs.
3) Develop algorithms for efficient decision-making and optimization, for individual IVs and a network of IVs.
4) Experimentation and validation of the developed methodologies to the physical testbed, a small-scale road network.
5) Support the ER to become a leading expert in the area of ITS and to transfer her acquired expertise to research institutions, companies and the public.
The project dealt at a theoretical level with uncertainty quantification and estimation of traffic variables. The proposed methods have shed light on the cutting-edge research areas of Bayesian inference, fault diagnosis and ITS. The main conclusions of the action are: (i) probabilistic traffic state estimation using data from IVs can reduce uncertainty up to 4 times compared to non-probabilistic methods, (ii) probabilistic traffic state estimation provides high-quality results even for low penetration rates of IVs, (iii) OD matrix estimation using measurements collected from IVs results in significantly better performance compared to measurements collected from fixed location sensors.
The overall scope of this project is to efficiently represent probabilistic uncertainties concerning IVs aiming to gain insights of their performance, reliability and safety. The main results of the project are:
• A novel Bayesian methodology for efficient estimation of the traffic density using data obtained by CAVs. Experimental studies were conducted involving both synthetic and real-world datasets.
• A novel methodology for spatial traffic demand estimation using traffic flow dynamics and link counts collected from a swarm of UAVs deployed over the network.
• A novel probabilistic traffic density estimation method utilising measurements collected from a swarm of UAVs deployed over the network.
The BITS project has so far delivered several scientific publications as follows (these are open access or will be when published).
Published articles:
• Englezou, Y., Timotheou, S. and Panayiotou, C. G. (2021) “Estimating the Origin-Destination Matrix using link count observations from Unmanned Aerial Vehicles,” Proceedings of the 2021 IEEE International Intelligent Transportation Systems Conference, pp. 3539-3544
• Kyriacou, V., Englezou, Y., Panayiotou, C. G., Timotheou, S. (2021). Estimating the posterior predictive distribution of the traffic density in multi-lane highways using spacing measurements,” Proceedings of the 2021 IEEE Intelligent Transportation Systems Conference, pp. 3634-3639
• Englezou, Y., Timotheou, S. and Panayiotou, C. G. (2022) “Probabilistic traffic density estimation using measurements from Unmanned Aerial Vehicles”, Proceedings of the 2022 IEEE International Conference on Unmanned Aircraft Systems, pp. 1381-1388
• Englezou, Y., Timotheou, S. and Panayiotou, C. G. (2022) “Simultaneous identification of sensor faults and origin-destination matrix estimation”, Proceedings of the 11th IFAC Symposium on Fault Detection, Supervision and Safety for Technical Processes, pp. 151-156
Articles accepted for publication:
• Englezou, Y., Timotheou, S. and Panayiotou, C. G. (2022) “Estimating the Origin-Destination Matrix using link count observations and traffic flow dynamics,” to appear in the Proceedings of the 2022 IEEE International Intelligent Transportation Systems Conference
Articles under review:
• Englezou, Y., Timotheou, S. and Panayiotou, G. C. (2022) “Static origin destination matrix estimation using macroscopic traffic flow dynamics”, under review at IEEE Transactions on Intelligent Transportation Systems (journal)
• Kyriacou, V., Englezou, Y., Panayiotou, G. C. and Timotheou, S. (2022) “Bayesian traffic state estimation using extended floating car data”, under review at IEEE Transactions on Intelligent Transportation Systems (journal)
• Sofokleous, M., Englezou, Y., Timotheou, S., Savva, A. and Panayiotou, G. C. “Network sensor location problem: a case study in the Republic of Cyprus” under review at Transportation Research Record
• Theocharides, K., Menelaou, C., Englezou, Y. and Timotheou, S. (2022) “Real-Time UAV Based Traffic State Estimation for a Multi-Regional Traffic Network” under review at Transportation Research Record
Articles in preparation:
• Englezou, Y., Timotheou, S. and Panayiotou, G. C. “Fault-tolerant Origin-Destination matrix estimation using network flow dynamics” (journal)
• Englezou, Y., Timotheou, S. and Panayiotou, G. C. “Utilising a swarm of Unmanned Aerial Vehicles for the efficient estimation of origin-destination matrices” (journal)
The Dissemination&Communication activities are:
- Press releases at the KIOS CoE and UCY newsletter and the KIOS CoE website
- Non-scientific publication: short video for #MyJobinResearch challenge published on REA Twitter account
- Social media dissemination through Twitter and Linkedin
- Dedicated website: https://www.kios.ucy.ac.cy/BITS/(se abrirá en una nueva ventana)
- The researcher was interviewed on national television and radio
- The work was presented at 5 international conferences
- The researcher participated at the Public discussion with the President of the European Commission accompanied by the President of the Republic of Cyprus at UCY, and several other visits at the KIOS CoE. The researcher was also featured in a special edition the ‘Green careers directory’ to showcase how people from across the globe are pursuing ‘Green careers’.
- Participated at the European Researcher's Night in 2020-2021
• A novel Bayesian methodology for efficient estimation of the traffic density using data obtained by CAVs. Experimental studies were conducted involving both synthetic and real-world datasets.
• A novel methodology for spatial traffic demand estimation using traffic flow dynamics and link counts collected from a swarm of UAVs deployed over the network.
• A novel probabilistic traffic density estimation method utilising measurements collected from a swarm of UAVs deployed over the network.
The BITS project has so far delivered several scientific publications as follows (these are open access or will be when published).
Published articles:
• Englezou, Y., Timotheou, S. and Panayiotou, C. G. (2021) “Estimating the Origin-Destination Matrix using link count observations from Unmanned Aerial Vehicles,” Proceedings of the 2021 IEEE International Intelligent Transportation Systems Conference, pp. 3539-3544
• Kyriacou, V., Englezou, Y., Panayiotou, C. G., Timotheou, S. (2021). Estimating the posterior predictive distribution of the traffic density in multi-lane highways using spacing measurements,” Proceedings of the 2021 IEEE Intelligent Transportation Systems Conference, pp. 3634-3639
• Englezou, Y., Timotheou, S. and Panayiotou, C. G. (2022) “Probabilistic traffic density estimation using measurements from Unmanned Aerial Vehicles”, Proceedings of the 2022 IEEE International Conference on Unmanned Aircraft Systems, pp. 1381-1388
• Englezou, Y., Timotheou, S. and Panayiotou, C. G. (2022) “Simultaneous identification of sensor faults and origin-destination matrix estimation”, Proceedings of the 11th IFAC Symposium on Fault Detection, Supervision and Safety for Technical Processes, pp. 151-156
Articles accepted for publication:
• Englezou, Y., Timotheou, S. and Panayiotou, C. G. (2022) “Estimating the Origin-Destination Matrix using link count observations and traffic flow dynamics,” to appear in the Proceedings of the 2022 IEEE International Intelligent Transportation Systems Conference
Articles under review:
• Englezou, Y., Timotheou, S. and Panayiotou, G. C. (2022) “Static origin destination matrix estimation using macroscopic traffic flow dynamics”, under review at IEEE Transactions on Intelligent Transportation Systems (journal)
• Kyriacou, V., Englezou, Y., Panayiotou, G. C. and Timotheou, S. (2022) “Bayesian traffic state estimation using extended floating car data”, under review at IEEE Transactions on Intelligent Transportation Systems (journal)
• Sofokleous, M., Englezou, Y., Timotheou, S., Savva, A. and Panayiotou, G. C. “Network sensor location problem: a case study in the Republic of Cyprus” under review at Transportation Research Record
• Theocharides, K., Menelaou, C., Englezou, Y. and Timotheou, S. (2022) “Real-Time UAV Based Traffic State Estimation for a Multi-Regional Traffic Network” under review at Transportation Research Record
Articles in preparation:
• Englezou, Y., Timotheou, S. and Panayiotou, G. C. “Fault-tolerant Origin-Destination matrix estimation using network flow dynamics” (journal)
• Englezou, Y., Timotheou, S. and Panayiotou, G. C. “Utilising a swarm of Unmanned Aerial Vehicles for the efficient estimation of origin-destination matrices” (journal)
The Dissemination&Communication activities are:
- Press releases at the KIOS CoE and UCY newsletter and the KIOS CoE website
- Non-scientific publication: short video for #MyJobinResearch challenge published on REA Twitter account
- Social media dissemination through Twitter and Linkedin
- Dedicated website: https://www.kios.ucy.ac.cy/BITS/(se abrirá en una nueva ventana)
- The researcher was interviewed on national television and radio
- The work was presented at 5 international conferences
- The researcher participated at the Public discussion with the President of the European Commission accompanied by the President of the Republic of Cyprus at UCY, and several other visits at the KIOS CoE. The researcher was also featured in a special edition the ‘Green careers directory’ to showcase how people from across the globe are pursuing ‘Green careers’.
- Participated at the European Researcher's Night in 2020-2021
Contributions to the state-of-the-art:
• In contrast with most literature methodologies that consider average static flows, we utilise traffic counts in small intervals and employ traffic dynamics to capture the dynamical nature of the system.
• Integration of traffic dynamics and fine-grained uncertain measurements of the demand estimation problem.
• Incorporation of traffic flow dynamics and link count observations collected from a swarm of UAVs.
• Investigation of the effect of spatial diversity of measurements for an individualized dynamical system.
• A novel probabilistic traffic density estimation method utilising measurements collected from a swarm of UAVs.
The proposed methods have high-quality estimation capabilities and allow for accurate predictions of the system. In addition, BITS algorithms yield efficient traffic state and demand estimation that allows the implementation of intelligent control policies that can reduce congestion and improve safety. The probabilistic uncertainty quantification for IVs is a step to the right direction to make such technologies part of our everyday lives. Recent studies have showed that an IV fleet will have a beneficial effect on noise pollution, leading to a general reduction of noise emissions, and their use can help reduce the significant environmental impacts of the transportation system in many ways. The project’s research and innovation can boost the interest of national companies, ministries or organisations, thus strengthening their competitiveness and growth in the market and establishing long-term collaborations with them.
• In contrast with most literature methodologies that consider average static flows, we utilise traffic counts in small intervals and employ traffic dynamics to capture the dynamical nature of the system.
• Integration of traffic dynamics and fine-grained uncertain measurements of the demand estimation problem.
• Incorporation of traffic flow dynamics and link count observations collected from a swarm of UAVs.
• Investigation of the effect of spatial diversity of measurements for an individualized dynamical system.
• A novel probabilistic traffic density estimation method utilising measurements collected from a swarm of UAVs.
The proposed methods have high-quality estimation capabilities and allow for accurate predictions of the system. In addition, BITS algorithms yield efficient traffic state and demand estimation that allows the implementation of intelligent control policies that can reduce congestion and improve safety. The probabilistic uncertainty quantification for IVs is a step to the right direction to make such technologies part of our everyday lives. Recent studies have showed that an IV fleet will have a beneficial effect on noise pollution, leading to a general reduction of noise emissions, and their use can help reduce the significant environmental impacts of the transportation system in many ways. The project’s research and innovation can boost the interest of national companies, ministries or organisations, thus strengthening their competitiveness and growth in the market and establishing long-term collaborations with them.