3L4AV

3L4AV logo Lifelong learning of dynamic objects detection and tracking in adverse conditions for autonomous vehicles

News

[31 Dec 2019]: The project is successfully concluded and all predetermined goals have been achieved!

[11 Dec 2019]: CTU team began a research visit at UTBM!

[6 Sep 2019]: Filip Majer presented our work at ECMR!

[2 Sep 2019]: UTBM team began a research visit at CTU!

[8 Aug 2019]: Dr. Zhi Yan gave a talk at CTU!

[21 Jun 2019]: Two papers, respectively from Tomas Vintr and Filip Majer, got accepted by ECMR’19, congratulations!

[13 Jun 2019]: Dr. Tomas Krajnik gave a talk at ETH Zurich!

[12 Jun 2019]: Dr. Zhi Yan gave a talk at ROSCon Fr 2019!

[10 Jun 2019]: CTU team began a research visit at UTBM!

[26 May 2019]: CTU team began a research visit at UTBM!

[19 May 2019]: George Broughton joined UTBM for a one-month research stay!

[3 May 2019]: The project wiki is online: https://github.com/gestom/3L4AVp/wiki

[5 Apr 2019]: Dr. Zhi Yan gave a talk at Journée GDR ISIS & GDR Robotique!

[1 Apr 2019]: UTBM team began a research visit at CTU!

[26 Jan 2019]: Tomas Vintr’s paper got accepted by ICRA’19, congratulations!

[15 Jan 2019]: CTU team began a research visit at UTBM!

[25 Dec 2018]: UTBM team began a research visit at CTU!

[14 Nov 2018]: Dr. Tomas Krajnik gave a great talk on his amazing FreMEn method!

[11 Nov 2018]: CTU team began a research visit at UTBM!

[8 Nov 2018]: UTBM team began a research visit at CTU!

[18 Aug 2018]: Filip Majer joined UTBM as an Erasmus student!

[20 Jul 2018]: A video of our dataset is online: https://youtu.be/d8TNc3vVMLA

[28 Jan 2018]: The project homepage is online!

[23 Jan 2018]: The project is kicked off!

About

3L4AV is a mobility research project between the Czech Technical University in Prague (CTU) in Czechia and the University of Technology of Belfort-Montbéliard (UTBM) in France. The research goal is to provide new methods for autonomous vehicles to improve the robustness of their perception system, in particular for dynamic objects detection and tracking in adverse conditions, such as during adverse weather and dense traffic.

In this project, we will investigate machine learning methods for multisensor systems deployed in autonomous vehicles. The heterogeneous nature of the sensory data will allow mutual training of the methods of dynamic object detection and tracking. On-the-fly, lifelong learning of the objects models for detection and tracking will be achieved through exploitation of the heterogeneity and amounts of data gathered by the vehicle’s sensors over long periods of time. In contrast to existing technologies which are mainly based on static models, our project will focus on development of adaptive models that are completed and refined based on the data gathered over long-time operation of the autonomous vehicle.

Partners

CTU

Tomas Krajnik (PI)
Assistant Professor
Michal Cap
Postdoc
Tomas Vintr
Ph.D. Student
Filip Majer
M.S. Student
Eliska Dvorakova
M.S. Student
Lucie Halodova
M.S. Student

UTBM

Zhi Yan (PI)
Assistant Professor
Yassine Ruichek (COORD)
Professor

Publications

  1. Zhi Yan, Li Sun, Tomas Krajnik, and Yassine Ruichek. EU Long-term Dataset with Multiple Sensors for Autonomous Driving. Preprint 2019.

  2. Filip Majer, Zhi Yan, George Broughton, Yassine Ruichek, and Tomas Krajnik. Learning to See Through Haze: Radar-based Human Detection for Adverse Weather Conditions. In Proceedings of the 2019 European Conference on Mobile Robots (ECMR), Prague, Czech Republic, September 2019.

  3. Tomas Vintr, Sergi Molina, Ransalu Senanayake, George Broughton, Zhi Yan, Jiri Ulrich, Tomasz Piotr Kucner, Chittaranjan Srinivas Swaminathan, Filip Majer, Maria Stachova, Achim J. Lilienthal, and Tomas Krajnik. Time-varying Pedestrian Flow Models for Service Robots. In Proceedings of the 2019 European Conference on Mobile Robots (ECMR), Prague, Czech Republic, September 2019.

  4. Tomas Vintr, Sergi Molina, Ransalu Senanayake, George Broughton, Zhi Yan, Jiri Ulrich, Tomasz Piotr Kucner, Chittaranjan Srinivas Swaminathan, Filip Majer, Maria Stachova, Achim J. Lilienthal, and Tomas Krajnik. Spatio-temporal Representation of Time-varying Pedestrian Flows. ICRA Workshop on Long-term Human Motion Prediction, Montreal, Canada, May 2019.

  5. Tomas Vintr, Zhi Yan, Tom Duckett, and Tomas Krajnik. Spatio-temporal Representation for Long-term Anticipation of Human Presence in Service Robotics. In Proceedings of the 2019 IEEE International Conference on Robotics and Automation (ICRA), Montreal, Canada, May 2019.

  6. Tomas Vintr, Kerem Eyisoy, Vanda Vintrova, Zhi Yan, Yassine Ruichek, and Tomas Krajnik. Spatiotemporal Models of Human Activity for Robotic Patrolling. In Proceedings of the 5th International Conference on Modelling and Simulation for Autonomous Systesm (MESAS), Prague, Czech Republic, October 2018.

Results


CZ MSMT project (No. FR-8J18FR018), PHC Barrande project (No. 40682ZH), 2018-2019 (2 years).

ctu logo utbm logo