Extension of vehicle behavior clustering to pedestrians and cyclists

Inria
April 30, 2023
Contact:N/A
Offerd Salary:Negotiation
Location:N/A
Working address:N/A
Contract Type:Other
Working Time:Negotigation
Working type:N/A
Ref info:N/A

2023-05835 - Extension of vehicle behavior clustering to pedestrians and cyclists

Contract type : Internship

Level of qualifications required : Master's or equivalent

Fonction : Internship Research

Context

The team-project ASTRA is continuation of the project RITS, which were a multidisciplinary project working on robotics for intelligent transportation systems. ASTRA is a partner team with VALEO to research and develop technologies for automated vehicles, from perception to motion planning and decision making.

Assignment

Context:

One of the challenges when testing automated vehicles in simulation is to propose a realistic environment with road users that are independent from the simulation itself and interact with each other. To built such agents two approaches can be taken: one based on knowledge about the environment itself or another based on the data obtained through observation, or both.

But first the data obtained needs to be processed before being used in a data- driven approach. For the observation in intersections made on 1, one needs to cluster the trajectories, with different sizes, into maneuvers (as seen in figure ref{fig:1}), to then classify each behavior existent in a maneuver according to its longitudinal velocity and acceleration (figures ref{fig:2} and ref{fig:3}). For vehicles, this was done applying the k-means algorithm twice, once for the position trace of the trajectory and another, on each detected cluster, on the longitudinal velocity and acceleration.

The distance measure used, since the data points do not have the same size, was the dynamic time warp 2, with a custom initialization function for the kmeans algorithm. There are multiple ways to do it, from using the longest common sub-sequence (LCSS) 3 to using a trained descriptor 4. Currently, the DTW distance measure is one of the best methods (if not the best) to align time series. Then, to define the number of cluster that exists in the data, the gap statistic was used 5, 6.

References:

1 J. Bock, R. Krajewski, T. Moers, S. Runde, L. Vater, and L. Eckstein, “The ind dataset: A drone dataset of naturalistic road user trajectories at german intersections,” in 2020 IEEE Intelligent Vehicles Symposium (IV). IEEE, 2020, pp. 1929–1934.

2 C. A. Ratanamahatana and E. Keogh, “Everything you know about dynamic time warping is wrong,” in Third workshop on mining temporal and sequential data, vol. 32. Citeseer, 2004.

3 M. Y. Choong, L. Angeline, R. K. Yin Chin, K. Beng Yeo, and K. T. Kin Teo, “Modeling of vehicle trajectory clustering based on lcss for traffic pattern extraction,” in 2017 IEEE 2nd International Conference on Automatic Control and Intelligent Systems (I2CACIS). IEEE, 2017, pp. 74–79.

4 S. Choi, J. Kim, and H. Yeo, “Trajgail: Generating urban vehicle trajectories using generative adversarial imitation learning,” Transporta- tion Research Part C: Emerging Technologies, vol. 128, p. 103091, 2021. Online. Available: https: // www. sciencedirect.com/science/article/pii/ S0968090X21001121

5 R. G. Ribeiro and R. Rios, “Temporal gap statistic: A new internal index to validate time series clustering,” Chaos, Solitons & Fractals, vol. 142, p. 110326, 2021. Online. Available: https: // www. sciencedirect.com/science/ article/pii/S0960077920307219

6 R. Tibshirani, G. Walther, and T. Hastie, “Estimating the number of clusters in a data set via the gap statistic,” Journal of the Royal Statistical Society: Series B (Statistical Methodology), vol. 63, no. 2, pp. 411–423, 2001. Online. Available: https: // rss.onlinel

Main activities

We are looking for one motivated M1 student to accomplish the following objectives:

  • Apply the same procedures described above to pedestrians and cyclists, making the necessary modifications given the different nature in comparison with vehicles.
  • Start the training of a simulation agent with the obtained information.
  • Depending on the quality of the results, publication on conference/journal may be considered.
  • Skills
  • Undergraduate on Computer Science or closely related fields.
  • Excellent programming skills in C++ and python.
  • Background in machine learning (reinforcement learning), simulation (Carla) and robotics; Knowledge of docker is a plus.
  • Proficiency in written and spoken English language (not a hard constraint).
  • General Information
  • Theme/Domain : Algorithmics, Computer Algebra and Cryptology Statistics (Big data) (BAP E)

  • Town/city : Paris

  • Inria Center : Centre Inria de Paris
  • Starting date : 2023-05-15
  • Duration of contract : 3 months
  • Deadline to apply : 2023-04-30
  • Contacts
  • Inria Team : ASTRA
  • Recruiter : De Moura Martins Gomes Nelson / [email protected]
  • The keys to success

    There you can provide a "broad outline" of the collaborator you are looking for what you consider to be necessary and sufficient, and which may combine :

  • tastes and appetencies,
  • area of excellence,
  • personality or character traits,
  • cross-disciplinary knowledge and expertise...
  • This section enables the more formal list of skills to be completed and 'lightened' (reduced) :

  • "Essential qualities in order to fulfil this assignment are feeling at ease in an environment of scientific dynamics and wanting to learn and listen."
  • " Passionate about innovation, with expertise in Ruby on Rails development and strong influencing skills. A thesis in the field of is a real asset."
  • About Inria

    Inria is the French national research institute dedicated to digital science and technology. It employs 2,600 people. Its 200 agile project teams, generally run jointly with academic partners, include more than 3,500 scientists and engineers working to meet the challenges of digital technology, often at the interface with other disciplines. The Institute also employs numerous talents in over forty different professions. 900 research support staff contribute to the preparation and development of scientific and entrepreneurial projects that have a worldwide impact.

    Instruction to apply

    Defence Security : This position is likely to be situated in a restricted area (ZRR), as defined in Decree No. 2011-1425 relating to the protection of national scientific and technical potential (PPST).Authorisation to enter an area is granted by the director of the unit, following a favourable Ministerial decision, as defined in the decree of 3 July 2012 relating to the PPST. An unfavourable Ministerial decision in respect of a position situated in a ZRR would result in the cancellation of the appointment.

    Recruitment Policy : As part of its diversity policy, all Inria positions are accessible to people with disabilities.

    Warning : you must enter your e-mail address in order to save your application to Inria. Applications must be submitted online on the Inria website. Processing of applications sent from other channels is not guaranteed.

    From this employer

    Recent blogs

    Recent news