Project-team positioning

The team carries out research across the full (software) stack of an autonomous driving system. This goes from perception of the environment to motion forecasting and trajectory optimization. While most of the techniques that we develop can stand on their own and are typically tested on our experimental platforms, we slowly move towards their complete integration. This will result in a functioning autonomous driving software stack. Beyond that point, we envision to focus our research on developing new methods that improve the individual performance of each of the stack’s components, as well as the global performance. This is alike to how autonomous driving companies operate (e.g. Waymo, Cruise, Tesla). The research topics that we explore are addressed by a very long list of international research laboratories and private companies. However, in terms of scope, we find the research groups of Prof. Dr.-Ing. Christoph Stiller (KIT), Prof. Marcelo Ang (NUS), Prof. Daniela Rus (MIT) and Prof. Wolfram Burgard (TRI California) to be the closest to us; we have active interconnections with these research groups (e.g. in the scope of the IEEE RAS Technical Committee AGV-ITS we are co-chairing any of the related annual Workshops). At the national level, we are cooperating in the framework of several R&D projects with UTC, Institut Pascal Clermont-Ferrand, University Gustave Eiffel, and the Inria Teams ASTRA, Acentauri and Convecs. Our main research originality relies in the combination of Bayesian approaches, AI technologies, Human-Vehicle interactions models, Robust State Estimation formalism, and a mix of Real experiments and Formal methods for validating our technologies in complex real world dynamic scenarios.

Our research on visual-inertial sensor fusion is closely related to the research carried out by the group of Prof. Stergios Roumeliotis (University of Minnesota) and the group of Prof. Davide Scaramuzza (University of Zurich). Our originality in this context is the introduction of closed-form solutions instead of filter-based approaches.

Collaborations

• Main long-term industrial collaborations: IRT Nanoelec & CEA $\sim$10 years), Toyota Motor Europe ($\sim$12 years, R&D center Zaventen-Brussels), Renault ($\sim$10 years), SME Easymile ($\sim$3 years), Sumitomo Japan (started in 2019), Iveco (3 years in the framework of STAR project). These collaborations are funded either by Industry or by National, Regional or European projects. The main outputs are common publications, softwares development, patents or technological transfers. Two former doctoral students of our team were recruited some years ago by respectively TME Zaventen and Renault R&D centers.
• Main academic collaborations: Institut Pascal Clermont-Ferrand, University Compiegne (UTC), University Gustave Eiffel, several Inria Teams (ASTRA, Acentauri, Convecs, Rainbow). These collaborations are most of the time conducted in the scope of various National R&D projects (FUI, PIA, ANR, etc.). The main outputs are scientific and software exchanges. In addition, we had a fruitful collaboration with B. Mourrain from the Inria-Aromath team. The goal was to find the analytical solution of a polynomial equation system that fully characterizes the cooperative visual-inertial sensor fusion problem with 2 agents.
• International cooperation for scientific exchanges and IEEE scientific events organization, e.g. NUS & NTU Singapore, Peking University, TRI Mountain View, KIT Karlsruhe, Coimbra University. We have constantly interacted with Prof. D. Scaramuzza from the university of Zurich to make our findings in visual inertial sensor fusion more usable in a realistic context (e.g., for the autonomous navigation of drones).

Project-team positioning

In his reference book Planning algorithms56, S. LaValle discusses the different dimensions that make the motion-planning a complex problem, which are the number of robots, the obstacle regions, the uncertainty of perception and action, and the allowable velocities. In particular, it is emphasized that multiple robot planning in complex environments are NP-hard problems which implies that exact approaches have exponential time complexities in the worst case (unless P=NP). Moreover, dynamic and uncertain environments, as human-populated ones, expand this complexity. In this context, we aim at scaling up decision-making in human-populated environments and in multi-robot systems, while dealing with the intrinsic limits of the robots and machines (computation capacity, limited communication). To address these challenges, we explore new algorithms and AI architectures following three directions: online planning/self-organization for real-time decision/control, machine learning to adapt to the complexity of uncertain environments, and combinatorial optimization to deal with offline constrained problems. Combining these approaches, seeking to scale up them, and also evaluating them with real platforms are also elements of originality of our work.

We share these goals with other laboratories/teams in the world, such as the Mobile Robotics Laboratory and Intelligent Systems "IntRoLab" at Sherbrooke University (Montreal) led by Prof. F. Michaud, the Learning Agents Research Group (LARG) within the AI Lab at the University of Texas (Austin) led by Prof. P. Stone, the Autonomous Systems Lab at ETH Zurich (Switzerland) led by R. Siegwart, the Robotic Sensor Networks Lab at the University of Minnesota (USA) led by Prof. V. Isler, and the Autonomous Robots Lab at NTNU (Norway) to cite a few. At Inria, we share some of the objectives with the LARSEN team in Nancy (multi-agent machine learning, multi-robot planning) and the RAINBOW team in Rennes (multi-UAV planning). We have also collaborations with the ACENTAURY team, in Sophia Antipolis, about autonomous navigation among humans. In France, among other labs involved on similar subjects, we can cite the LAAS (CNRS, Toulouse), the ISIR lab (CNRS, Paris Sorbonne Univ.) and the MAD team in Caen University. In the more generic domain of problem solving, we can mention the TASC team at LS2N (Nantes), the CRIL lab (Lens) and, at the international level, we can cite the Insight Centre for Data Analytics (Cork, Ireland).

Collaborations

• Social navigation: collaboration with Julie Dugdale and Dominique Vaufreydaz, Asso. Prof. at LIG Laboratory and with Prof. Philippe Martinet from Acentauri Inria team (common PhD students and projects: ANR VALET, ANR HIANIC, ANR ANNAPOLIS).
• Multi-robot exploration: Prof. Cedric Pradalier from GeorgiaTech Metz/CNRS (France) leading the European project BugWright2 — Prof. Luca Schenato from University of Padova (Italy).
• Swarm of UAVs: Prof. Isabelle Guérin-Lassous from LIP lab/Lyon 1 (common PhD students and projects: ANR CONCERTO, INRIA/DGA DynaFlock) — Research Director Isabelle Fantoni from CNRS/LS2N Lab in Nantes (common projects: ANR CONCERTO, Equipex TIRREX, ANR VORTEX)
• Machine Learning: strong collaborations with Christian Wolf, Naver Labs Europe & LIRIS lab/INSA8 (common PhD students and projects: ANR DELICIO, ANR AI Chair REMEMBER) and with Laetitia Matignon, Asso. Prof. at LIRIS lab/Univ Lyon 1 (common projects and Phd/Master students, eg. BPI/ANR SOLAR-Nav) François Charpillet and Olivier Buffet, Inria researchers in LARSEN team, Nancy (common publications, projects surch as ANR PLASMA, BPI/ANR SOLAR-Nav).
• Semantic grids for autonomous navigation: Collaboration with Toyota Motor Europe (TME), R&D center Zaventen (Brussels). R&D contracts, common publications and patents.
• Constrained optimization: Omar Rifki (Post-doc) and Thierry Garaix (Ass. Prof) at LIMOS Ecole des Mines de Saint Etienne — Pascal Lafourcade and François Delobel, Asso. Prof. at LIMOS Université Clermont-Ferrand.

7.1.1 CMCDOT

• Functional Description:
  CMCDOT is a Bayesian filtering system for dynamic occupation grids, allowing parallel estimation of occupation probabilities for each cell of a grid, inference of velocities, prediction of the risk of collision and association of cells belonging to the same dynamic object. Last generation of a suite of Bayesian filtering methods developed in the Inria eMotion team, then in the Inria Chroma team (BOF, HSBOF, ...), it integrates the management of hybrid sampling methods (classical occupancy grids for static parts, particle sets for parts dynamics) into a Bayesian unified programming formalism , while incorporating elements resembling the Dempster-Shafer theory (state "unknown", allowing a focus of computing resources). It also offers a projection system of the estimated scene in the near future, to reference potential collisions with the ego-vehicle or any other element of the environment, as well as very low cost pre-segmentation of coherent dynamic spaces (taking into account speeds). It takes as input instantaneous occupation grids generated by sensor models for different sources, the system is composed of a ROS package, to manage the connectivity of I / O, which encapsulates the core of the embedded and optimized application on GPU Nvidia (Cuda), allowing real-time analysis of the direct environment on embedded boards (Tegra X1, X2). ROS (Robot Operating System) is a set of open source tools to develop software for robotics. Developed in an automotive setting, these techniques can be exploited in all areas of mobile robotics, and are particularly suited to highly dynamic and uncertain environment management (e.g. urban scenario, with pedestrians, cyclists, cars, buses, etc.).
• Contact:
  Christian Laugier

7.1.2 Ground Elevation and Occupancy Grid Estimator (GEOG - Estimator)

• Functional Description:
  GEOG-Estimator is a system of joint estimation of the shape of the ground, in the form of a Bayesian network of constrained elevation nodes, and the ground-obstacle classification of a pointcloud. Starting from an unclassified 3D pointcloud, it consists of a set of expectation-maximization methods computed in parallel on the network of elevation nodes, integrating the constraints of spatial continuity as well as the influence of 3D points, classified as ground-based or obstacles. Once the ground model is generated, the system can then construct a occupation grid, taking into account the classification of 3D points, and the actual height of these impacts. Mainly used with lidars (Velodyne64, Quanergy M8, IBEO Lux), the approach can be generalized to any type of sensor providing 3D pointclouds. On the other hand, in the case of lidars, free space information between the source and the 3D point can be integrated into the construction of the grid, as well as the height at which the laser passes through the area (taking into account the height of the laser in the sensor model). The areas of application of the system spread across all areas of mobile robotics, it is particularly suitable for unknown environments. GEOG-Estimator was originally developed to allow optimal integration of 3D sensors in systems using 2D occupancy grids, taking into account the orientation of sensors, and indefinite forms of grounds. The ground model generated can be used directly, whether for mapping or as a pre-calculation step for methods of obstacle recognition or classification. Designed to be effective (real-time) in the context of embedded applications, the entire system is implemented on Nvidia graphics card (in Cuda), and optimized for Tegra X2 embedded boards. To ease interconnections with the sensor outputs and other perception modules, the system is implemented using ROS (Robot Operating System), a set of opensource tools for robotics.
• Contact:
  Christian Laugier

7.1.3 Zoe Simulation

• Name:
  Simulation of INRIA's Renault Zoe in Gazebo environment
• Keyword:
  Simulation
• Functional Description:
  This simulation represents the Renault Zoe vehicle considering the realistic physical phenomena (friction, sliding, inertia, ...). The simulated vehicle embeds sensors similar to the ones of the actual vehicle. They provide measurement data under the same format. Moreover the software input/output are identical to the vehicle's. Therefore any program executed on the vehicle can be used with the simulation and reciprocally.
• Contact:
  Christian Laugier

7.1.4 Hybrid-state E*

• Name:
  Path planning with Hybrid-state E*
• Keywords:
  Planning, Robotics
• Functional Description:

  Considering a vehicle with the kinematic constraints of a car and an environment which is represented by a probabilistic occupancy grid, this software produces a path from the initial position of the vehicle to its destination. The computed path may include, if necessary, complex maneuvers. However the suggested path is often the simpler and the shorter.

  This software is designed to take benefit from bayesian occupancy grids such as the ones computed by the CMCDOT software.

• URL:
  https://­team.­inria.­fr/­chroma/­en/
• Contact:
  Christian Laugier
• Partner:
  CEA

7.1.5 Pedsim_ros_AV

• Name:
  Pedsim_ros_AV
• Keywords:
  Simulator, Multi-agent, Crowd simulation, Autonomous Cars, Pedestrian
• Scientific Description:

  These ROS packages are useful to support robotic developments that require the simulation of pedestrians and an autonomous vehicle in various shared spaces scenarios. They allow: 1. in simulation, to pre-test autonomous vehicle navigation algorithms in various crowd scenarios, 2. in real crowds, to help online prediction of pedestrian trajectories around the autonomous vehicle.

  Individual pedestrian model in shared space (perception, distraction, personal space, pedestrians standing, trip purpose). Model of pedestrians in social groups (couples, friends, colleagues, family). Autonomous car model. Pedestrian-autonomous car interaction model. Definition of shared space scenarios: 3 environments (business zone, campus, city centre) and 8 crowd configurations.

• Functional Description:
  Simulation of pedestrians and an autonomous vehicle in various shared space scenarios. Adaptation of the original Pedsim_ros model to simulate heterogeneous crowds in shared spaces (individuals, social groups, etc.). The car model is integrated into the simulator and the interactions between pedestrians and the autonomous vehicle are modeled. The autonomous vehicle can be controlled from inside the simulation or from outside the simulator by ROS commands.
• URL:
  https://­github.­com/­maprdhm/­pedsim_ros_AV
• Publications:
  hal-02194735, hal-02514963
• Contact:
  Manon Predhumeau
• Participants:
  Manon Predhumeau, Anne Spalanzani, Julie Dugdale, Lyuba Mancheva
• Partner:
  LIG

7.1.6 S-NAMO-SIM

• Name:
  S-NAMO Simulator
• Keywords:
  Simulation, Navigation, Robotics, Planning
• Functional Description:
  2D Simulator of NAMO algorithms (Navigation Among Movable Obstacles) and MR-NAMO (Multi-Robot NAMO), ROS compatible
• Release Contributions:
  Creation
• Contact:
  Olivier Simonin

7.1.7 SimuDronesGR

• Name:
  Simultion of UAV fleets with Gazebo/ROS
• Keywords:
  Robotics, Simulation
• Functional Description:
  The simulator includes the following functionality : 1) Simulation of the mechanical behavior of an Unmanned Aerial Vehicle : * Modeling of the body's aerodynamics with lift, drag and moment * Modeling of rotors' aerodynamics using the forces and moments' expressions from Philppe Martin's and Erwan Salaün's 2010 IEEE Conference on Robotics and Automation paper "The True Role of Accelerometer Feedback in Quadrotor Control". 2) Gives groundtruth informations : * Positions in East-North-Up reference frame * Linear velocity in East-North-Up and Front-Left-Up reference frames * Linear acceleration in East-North-Up and Front-Left-Up reference frames * Orientation from East-North-Up reference frame to Front-Left-Up reference frame (Quaternions) * Angular velocity of Front-Left-Up reference frame expressed in Front-Left-Up reference frame. 3) Simulation of the following sensors : * Inertial Measurement Unit with 9DoF (Accelerometer + Gyroscope + Orientation) * Barometer using an ISA model for the troposphere (valid up to 11km above Mean Sea Level) * Magnetometer with the earth magnetic field declination * GPS Antenna with a geodesic map projection.
• Release Contributions:
  Initial version
• Contact:
  Olivier Simonin
• Partner:
  Insa de Lyon

7.1.8 spank

• Name:
  Swarm Protocol And Navigation Kontrol
• Keyword:
  Protocoles
• Functional Description:
  Communication and distance measurement in an uav swarm
• URL:
  https://­gitlab.­inria.­fr/­dalu/­spank
• Contact:
  Stéphane D'alu
• Participant:
  Stéphane D'alu

7.2.1 Chroma Aerial Robots platform

Participants: Olivier Simonin, Alessandro Renzaglia, Johan Faure.

This platform is composed of :

• Four quadrirotor PX4 Vision UAVs (Unmaned Aerial Vehicles), acquired 4 in 2021 and 2022. This platform is funded and used in the projects "Dynaflock" (Inria-DGA) and ANR "CONCERTO" (the team also owns 5 Parrot Bebop UAVs).
• 10 Crazyflies mini-drones and a LightHouse localization system, acquired in 2024.
• Two outdoor inflatable aviaries of 6m(L) x 4m(l) x 5m(H) each.
• Two indoor inflatable aviaries of 5m(L) x 3m(l) x 2.5m(H) each.

7.2.2 Barakuda

Participants: Lukas Rummelhard, Thomas Genevois, Robin Baruffa, Hugo Bantignies, Nicolas Turro.

Within the scope of the ANR Challenge MOBILEX, a 4-wheel all-terrain Barakuda robot from Shark Robotics is provided, and will be equipped with sensors and compute capabilities to achieve increasing level of autonomy.

8.1.1 Evolution of the CMCDOT framework

Participants: Robin Baruffa, Hugo Bantignies, Nicolas Turro, Lukas Rummelhard.

Recognized as one of the core technologies developed within the team over the years (see related sections in previous activity report of Chroma, and previously e-Motion reports), the CMCDOT framework is a generic Bayesian Perception framework, designed to estimate a dense representation of dynamic environments and the associated risks of collision, by fusing and filtering multi-sensor data. This whole perception system has been developed, implemented and tested on embedded devices, incorporating over time new key modules. Now extended to 3D for grid fusion and incorporating semantic states in the whole process, this framework, and the corresponding software, is at the core of many important industrial partnerships and academic contributions, and to be the subject of important developments, both in terms of research and engineering. Mainly focused this year on engineering issues, the activity on the important software and technological base has allowed a much stronger modularity and robustness, integrating into our solutions the latest reference developments in the mobile robotics community :

• Migration from ROS1 to ROS2 : involving the porting of all the developed perception and navigation software, and all the embedded GPU memory transfer optimizations, the migration was finally decided and carried out this year.
• Integration in the Nav2 framework : to benefit from its modularity and flexibility, allowing to develop specific bricks progressively while benefiting from a set of proven generic functionalities, the perception system has been interfaced with and the navigation system integrated in the Nav2 framework.
• Overall Software stability : synergistically with maturation efforts to spin off a start-up based on these technologies, huge steps to industrialize the software have been taken.

8.1.2 Validation of AI-based algorithms in autonomous vehicles

Participants: Jean-Baptiste Horel, Alessandro Renzaglia, Khushdeep Mann, Radu Mateescu [Convecs], Wendelin Serwe [Convecs], Christian Laugier.

In the last years, there has been an increasing demand for regulating and validating intelligent vehicle systems to ensure their correct functioning and build public trust in their use. Yet, the analysis of safety and reliability poses a significant challenge. More and more solutions for autonomous driving are based on complex AI-based algorithms whose validation is particularly challenging to achieve. An important part of our work has been devoted to tackle this problem, finding new approaches to validate probabilistic perception algorithms for autonomous driving, and investigating how simulations can be combined with real experiments via our framework of augmented reality to solve this challenge. This activity, started with our participation in the European project Enable-S3 (2016-2019), has then continued with the PRISSMA project (2021-2024) where we participated in collaboration with the Inria CONVECS team. This project, funded by the French government in the framework of the “Grand Défi: Sécuriser, certifier et fiabiliser les systèmes fondés sur l'intelligence artificielle”, regrouped several companies and public institutes to tackle the challenges of the validation and certification of artificial intelligence based solutions for autonomous mobility.

Within this context, our main effort has been focused on proposing a new approach that, given a high-level and human-interpretable description of a critical situation, generates relevant AV scenarios and uses them for automatic verification. To achieve this goal, we integrate two recently proposed methods for the generation and the verification that are based on formal verification tools. First, we proposed the use of formal conformance test generation tools to derive, from a verified formal model, sets of scenarios to be run in a simulator 54. Second, we model check the traces of the simulation runs to validate the probabilistic estimation of collision risks. Using formal methods brings the combined advantages of an increased confidence in the correct representation of the chosen configuration (temporal logic verification), a guarantee of the coverage and relevance of automatically generated scenarios (conformance testing), and an automatic quantitative analysis of the test execution (verification and statistical analysis on traces) 55. The last part of the project in 2024 has been mainly dedicated to carry out an experimental campaign at the Transpolis testing facility to showcase the main results of this work 25.

8.2.1 Situation Understanding and Motion Forecasting

Participants: Kaushik Bhowmik, Anne Spalanzani.

Forecasting the motion of surrounding traffic is one of the key challenges in the quest to achieve safe autonomous driving technology. Current state-of-the-art deep forecasting architectures are capable of producing impressive results. However, in many cases, they also output completely unreasonable trajectories, making them unsuitable for deployment. In 2023 we started to work on predicting the motion of heterogeneous agents in urban environment (K. Bhowmik's PhD). The goal is to predict the behavior of pedestrians, cyclists, and even electrical scooters. Using Deeplearning methods a existing big datasets of urban scenes, we plan to build robusts models of various agents. When their motion is highly unpredictable, we aim at warning the autonomous vehicle that a specific zone around it could be dangerous.

8.2.2 Cross-dataset evaluation of Semantic Grid Generation Models

Participants: Manuel Alejandro Diaz-Zapata [PhD. student], Wenqian Liu, Robin Baruffa, Christian Laugier, Jilles Dibangoye, Olivier Simonin.

A comprehensive cross-dataset evaluation framework for Bird's-Eye View (BEV) semantic segmentation was introduced in 2024 through the BEVal study 23. This work addresses a critical gap in BEV segmentation research, which historically focused on single-dataset evaluation (typically nuScenes), potentially leading to overly specialized models that may fail when faced with different environments or sensor configurations. The study evaluated three state-of-the-art BEV segmentation approaches across different sensor modalities: LSS (camera-only), LAPT (early camera-LiDAR fusion), and LAPT-PP (late camera-LiDAR fusion). Testing was conducted across the nuScenes and Woven Planet datasets, examining performance on three key semantic categories: vehicles, humans, and drivable areas, (see Fig. 3). A notable finding was that models heavily dependent on LiDAR data, particularly LAPT-PP, showed significant performance degradation in cross-dataset scenarios, with up to 80% reduction in IoU scores. In contrast, camera-based models demonstrated better generalization capabilities across datasets.

The research also explored multi-dataset training, combining both datasets during the training phase, which showed promising results with models achieving more balanced performance across both datasets. Particularly noteworthy was the improvement in human segmentation performance on the Woven Planet dataset, which saw a 10-15% increase when trained with the more diverse nuScenes data included. The study's findings underscore the importance of cross-dataset validation in developing robust BEV segmentation models for autonomous driving applications. The results suggest that while sensor fusion approaches can achieve high performance on individual datasets, their generalization capabilities may be limited by dataset-specific characteristics, particularly in LiDAR configurations. This highlights the need for diverse training data and robust validation strategies in developing reliable autonomous driving systems.

Show image description

Qualitative Results for Cross-dataset Validation

8.2.3 AI-driven Motion Planning and Temporal Multi-Modality Sensor Fusion for Semantic Grid Prediction

Participants: Gustavo Salazar-Gomez, Manuel Alejandro Diaz-Zapata, Wenqian Liu, Christian Laugier, Anne Spalanzani, Khushdeep Singh Mann.

In autonomous driving, accurately perceiving the environment despite occlusions is vital for safe and reliable navigation. TLCFuse was designed to enhance occlusion-aware semantic segmentation in autonomous driving scenarios. By integrating temporal cues and fusing multi-modal inputs from LiDAR, cameras, and ego-motion, TLCFuse creates robust bird's-eye-view (BEV) semantic grids capable of detecting occluded objects. The model processes temporal multi-step inputs using a transformer-based encoder-decoder architecture to extract spatio-temporal features effectively. Evaluated on the nuScenes dataset, TLCFuse outperforms state-of-the-art methods, particularly in scenarios with occlusions. The model's adaptability is further demonstrated in downstream tasks, including multi-step BEV prediction and ego-vehicle trajectory forecasting, showcasing its potential for broader applications in autonomous driving systems. This work was presented at IEEE Intelligent Vehicle(IV) 2024 30.

In the context of our cooperation with Toyota Motor Europe, we focus on developing advanced and reliable approaches for autonomous driving capabilities, promoting ethical and explainable decision making.

Working on AI-driven motion planning for autonomous vehicles was the main focus in 2024 with the start of Gustavo Salazar's PhD. The main objective is to propose an AI-driven framework for Trajectory Planning & Decision-making for autonomous driving, which takes into account the knowledge provided by the perception and tracking modules and interactions with other agents to plan the optimal and safest trajectory, avoiding the obstacles and complying with the traffic rules.

8.2.4 Dynamic Scene Prediction for Urban Driving Scene Using Occupancy Grid Maps

Participants: Rabbia Asghar [PhD. student], Lukas Rummelhard, Anne Spalanzani, Christian Laugier.

Accurate prediction of driving scenes is essential for road safety and autonomous driving. This remains a challenging task due to uncertainty in sensor data, the complex behaviors of agents, and possibility of multiple feasible futures. Over the course of Rabbia's PhD, we have addressed the prediction problem by representing the scene as dynamic occupancy grid maps (DOGMs) and integrating semantic information. We employ deep learning-based spatiotemporal approach to predict the evolution of scene and learn complex behaviours.

Carrying forward the work, in 2024, we consider prediction of flow or velocity vectors in the scene. We propose a novel multi-task framework that leverages dynamic OGMs and semantic information to predict both future vehicle semantic grids and the future flow of the scene. This incorporation of semantic flow not only offers intermediate scene features but also enables the generation of warped semantic grids. Evaluation on the real-world NuScenes dataset demonstrates improved prediction capabilities and enhanced ability of the model to retain dynamic vehicles within the scene. This work was published at IEEE ITSC 2024 21.

Developing further on these results, a novel approach to bring together agent-agnostic scene predictions and agent-specific predictions is proposed. This work not only focuses on predicting the behaviors of vehicle agents but also identifies other dynamic entities within the scene and anticipates their possible evolutions. We predict scene evolution as agent semantic grids, dynamic occupancy grid maps, and scene flow grids. Leveraging flow-guided prediction, these grids enable diverse future predictions for vehicles and dynamic elements while also capturing inter-grid dependencies through specialized loss functions. Evaluations on real-world NuScenes and Woven Planet datasets demonstrate superior prediction performances for dynamic vehicles and generic dynamic scene elements compared to baseline method (see Fig. 4). This work was presented at a poster session at IEEE IROS 2024 41.

Show image description

Complex driving scene forcasting

8.4.1 Motion-planning in dense pedestrian environments

Participants: Thomas Genevois, Anne Spalanzani.

Under the coordination of Anne Spalanzani, we study new motion planning algorithms to allow robots/vehicles to navigate in human populated environment while predicting pedestrians' motions. We model pedestrians and crowd behaviors using notions of social forces and cooperative behavior estimations. We investigate new methods to build safe and socially compliant trajectories using theses models of behaviors. We propose proactive navigation solutions as well as deep learning ones.

since 2018 we've been working on modelling crowds and autonomous vehicles using Extended Social Force Models (in the scope of Manon Predhumeau's PhD) and Proactive Navigation for navigating dense human populated environments (in the scope of Maria Kabtoul's PhD). The focus of the first work has been on the realistic simulation of crowds in shared spaces with an autonomous vehicle (AV), see Fig. 5.a. We proposed an agent-based model for pedestrian reactions to an AV in a shared space, based on empirical studies and the state of the art. The model includes an AV with Renault Zoé car's characteristics, and pedestrians' reactions to the vehicle. The model extends the Social Force Model with a new decision model, which integrates various observed reactions of pedestrians and pedestrians groups. The SPACiSS simulator is an opensource simulator still used by the team. The focus of the second work was a pedestrian-vehicle interaction behavioral model and a proactive navigation algorithm. The model estimates the pedestrian's cooperation with the vehicle in an interaction scenario by a quantitative time-varying function. Using this cooperation estimation the pedestrian's trajectory is predicted by a cooperation-based trajectory planning model. We then used this cooperation based behavioral model to design a proactive longitudinal velocity controller.

In 2023, Thomas Genevois worked on improving the previous works done in this topic. He proposed some improvement on the SPACiSS simulator, proposed a version of a huamn and interaction aware collision avoidance that can be used in the global architecture of the zoe car. He then linked the SPACiSS simulator to his Augmented Reality simulator so that the navigation strategies among crowds can be tested using the real zoé car 52. His PhD has been defended in June 2024.

Show image description

A scene of an Autonomous Vehicle with pedestrians

Show image description

A scene of an Autonomous Vehicle with pedestrians

8.4.2 Attention Graph for Multi-Robot Social Navigation with Deep Reinforcement Learning

Participants: Jacques Saraydaryan, Laetitia Matignon (Lyon1/LIRIS), Olivier Simonin.

Show image description

Multisoc architecture

Learning robot navigation strategies among pedestrian is crucial for domain based applications. Combining perception, planning and prediction allows to model the interactions between robots and pedestrians, resulting in impressive outcomes especially with recent approaches based on deep reinforcement learning (RL). However, these works do not consider multi-robot scenarios.

In this work, we define MultiSoc, a new method for learning multi-agent socially aware navigation strategies using RL (see illustration fig. 6). Inspired by recent works on multi-agent deep RL, our method leverages graph-based representation of agent interactions, combining the positions and fields of view of entities (pedestrians and agents). Each agent uses a model based on two Graph Neural Network combined with attention mechanisms. First an edge-selector produces a sparse graph, then a crowd coordinator applies node attention to produce a graph representing the influence of each entity on the others. This is incorporated into a model-free RL framework to learn multi-agent policies. We evaluated our approach on simulation and provide a series of experiments in a set of various conditions (number of agents/pedestrians). Empirical results show that our method learns faster than social navigation deep RL mono-agent techniques, and enables efficient multi-agent implicit coordination in challenging crowd navigation with multiple heterogeneous humans. Furthermore, by incorporating customizable meta-parameters, we can adjust the neighborhood density to take into account in our navigation strategy. This work has been published in AAMAS 2024 conference 24. We now continue this work in the postdoc supported by the SOLAR-Nav project starting in 2025.

8.4.3 Deep-Learning for vision-based navigation

Participants: Pierre Marza (PhD. student CITI & LIRIS), Christian Wolf (NaverLabs europe), Olivier Simonin, Laetitia Matignon (LIRIS/Lyon 1).

PhD thesis of Pierre Marza (2021-24). In the context of the REMEMBER AI Chair, Pierre Marza's PhD studied high-level navigation skills for autonomous agents in 3D environments.

In this work, we focuse on training neural agents able to perform semantic mapping, both from augmented supervision signal, and with proposed neural-based scene representations. Neural agents are often trained with Reinforcement Learning (RL) from a sparse reward signal. Guiding the learning of scene mapping abilities by augmenting the vanilla RL supervision signal with auxiliary spatial reasoning tasks to solve are shown to help with navigating efficiently. Instead of improving the training signal of neural agents, we showed how incorporating specific neural-based representations of semantics and geometry within the architecture of the agent can help improve performance in goal-driven navigation 58. Since 2023, we study how to best explore a 3D environment in order to build neural-based representations of space that are as satisfying as possible based on robotic-oriented metrics. This work has been published in IROS 2024 conference 26. Finally, we moved from navigation-only agents trained to generalize to new environments to multi-task agents. See Publication in CVPR 2024 27. Pierre Marza defended his PhD thesis on November 25th, 2024 35.

8.5.1 Navigation Among Movable Obstacles (NAMO): extension to multi-robot and social constraints

Participants: Jacques Saraydaryan, Olivier Simonin, David Brown (Ing. Inria).

Show image description

S-NAMO paths and Pepper robot

Show image description

S-NAMO paths and Pepper robot

In Chroma, we study NAMO problems since 2018 (Navigation Among Movable Obstacles). The PhD thesis of Benoit Renault (2019-23) allowed us to extend existing algorithms with the social cost of object placement (Social-NAMO 62) and the generalisation to multi-robot NAMO problems.

In 2023, we obtained an Inria ADT project, called NAMOEX, aiming to extend the Benoit Renault's NAMO simulator. By hiring David Brown (Ing.) in NAMOEX, we robustified the simulator and studied multi-robot NAMO strategies, leading to a publication at IROS 2024 conference 29. In 2024, we developped a connexion between our S-NAMO planner and the Gazebo robotic simulator, also extended as a monitoring tool for real experiments with a Turtlebot robot, see Fig. 7.

8.5.2 Multi-UAV Exploration and Visual Coverage of 3D Environments

Participants: Alessandro Renzaglia, Benjamin Sportich, Kenza Boubakri, Olivier Simonin.

Multi-robot teams, especially when they include Unmanned Aerial Vehicles (UAVs), are highly efficient systems for assisting humans in gathering information on large and complex environments. In these scenarios, cooperative coverage and exploration are two fundamental problems with numerous important applications, such as mapping, inspection of complex structures, and search and rescue missions. In both cases, the goal of the robots is to navigate through the environment and cooperate to acquire new information in order to complete the mission in the shortest possible time 64. Part of this activity was recently conducted in the framework of the European H2020 project BugWright2, which ended in early 2024, where a central use-case focused on corrosion detection and inspection on cargo ships. In this context, we investigated path planning strategies for fleets of UAVs to efficiently inspect large surfaces such as ship hulls and water tanks 53.

Show image description

Illustration of UAV trajectories during an exploration mission.

This line of research is now mainly carried out within the ANR project AVENUE, which focuses on the coordination of a fleet of UAVs to efficiently explore and reconstruct unknown 3D outdoor scenes. In particular, in this research work, both the task execution time and the accuracy of the final map are considered by proposing a distributed strategy where each robot can dynamically switch between two different behaviors: exploration mode and coverage mode. The exploration mode aims at obtaining as much and as fast as possible information about the structure of the environment and surfaces, even if imperfect, while the coverage mode aims at refining the mesh based on the collected information. Formally, the two behaviors are obtained by defining different information gain functions, both using a Euclidean Signed Distance Field (ESDF) based representation of the environment, to define future viewpoints. See illustration 8.

8.5.3 Multi-Robot Path Planning with Cable Constraints

Participants: Xiao Peng, Christine Solnon, Olivier Simonin, Simon Ferrier, Damien Rambeau.

This research was done within the context of the european project BugWright2 (2020-24) that aimed at designing an adaptable autonomous multi-robot solution for inspecting ship outer hulls.

PhD thesis of Xiao Peng. The aim of this PhD thesis was to design algorithms to plan the missions of a set of mobile robots, each attached to a flexible cable (a tether), while ensuring that the cables never cross. In 60, we formally defined the Non-Crossing MAPF (Multi-Agent Path Finding) problem and showed how to compute lower and upper bounds for this problem by solving well known assignment problems. We introduced a Variable Neighbourhood Search approach for improving the upper bound, and we introduced a CP model for solving the problem to optimality. In 59, we extended this approach to the case of non point-sized robots. In 2023/24, we extended the study to the coverage of an area with obstacles and fobiden arreas using a single tethered robot. Our solving approach relies on combining spanning-tree computation and ILP. This work has been accepted for publication in IEEE Robotics and Automation Letters (RAL journal) 17. Then we examined different algotihms to divide the environment in order to distribute the coverage to several robots. Xiao Peng defended his PhD thesis on February 19th, 2024 36.

Simulations and Robotics experimentations

In parallel of the thesis we developped an experimental setup in order to examine the complexity of deploying the proposed algorithms with real robots. First, Damien Rambeau (Ing.) developped a realistic simulation of crawler robots attached to tethers with Gazebo (he defined a model of cable able to enroll around obstacles). Second, Simon Ferrier (Eng.), developped a real robot demonstrator based on several turtlebot mobile robots that were attached to windable cables. At the end, we proposed an hybrid demonstrator were we merged real crawler robots (from CNRS/GeorgiaTech Metz partner) with simulated ones, in order to study complex multi-tethered-crawler scenarios. We realized a video and a presentation of this work for the ICRA 40th anniversary conference 43.

8.5.4 Multi-robot agricultural itineraries planning (NINSAR project)

Participants: Simon Ferrier, Olivier Simonin, Alessandro Renzaglia.

Show image description

(a) Rows allocation example (b) Solution given by our heuristic for a field of 50 rows covered by 5 robots.

Show image description

(a) Rows allocation example (b) Solution given by our heuristic for a field of 50 rows covered by 5 robots.

The NINSAR project, part of the PEPR AgroEcologie & Numerique, aims to use robots to help the agricultural sector. In the PhD thesis of Simon Ferrier, started on summer 2024, we define and study algorithms for the mission planning of a fleet of mobile robots. First work concerns the planning of itineries of each robot to treat all rows of a field in minimum time. We defined an optimization problem where the criteria are minimzing the makespan (mission duration) and the computational time.

To deal with this planning problem, we proposed two approaches: 1) a heuristic based on the division of the total sum of line lengths combined with a solution improvement algorithm 2) a Branch and Bound algorithm computing the optimal solution, and an anytime version. An illustration of the problem and of a solution are presented in figure 9. Results show that the heuristic approach is able to obtain very good results (close to the optimality) in a very short time while the exact approach presents a combinatory explosition with the number of robots deployed.

Next objectives are to integrate the execution of plans with unespected events requiring (local) online replanning.

8.5.5 Cooperative Gradient Estimation for Multi-Robot 3D Source Seeking

Participants: Alessandro Renzaglia.

As part of our research on online path planning and coordination of multi-robot systems, we recently began collaborating with the Gipsa-lab in Grenoble to study the problem of source seeking 47, 63. This problem is widely studied in robotics due to its numerous applications, including search and rescue operations, environmental monitoring, and pollution source detection. In source-seeking scenarios, an unknown source generates a 3D spatially distributed signal field. A team of autonomous robots, equipped with sensors to measure signal intensity at their respective locations, aims to cooperatively localize the source by leveraging the information collected and shared within the team.

Our work focuses on developing novel strategies to estimate the gradient of the signal field cooperatively. This gradient estimation is then used to guide the multi-robot system toward the source location. A key contribution of our research has been the use of geometric optimization approaches, such as centroidal Voronoi partitions on 3D surfaces, to define suitable multi-robot formations and to then exploit their geometric properties to obtain analytically a accurate and robust gradient estimations 22. This class of solutions offers the advantages of low computational cost, distributed implementation for increased robustness, and suitability for online team reconfiguration. This research is still ongoing and we are actively improving the initial solutions. In particular, current investigations focus on the impact of heterogeneous sensors with varying noise levels and how the team can adapt to such data to maintain robust and reliable gradient estimations.

8.6.1 Communications in swarm of UAVs

Participants: Théotime Balaguer (PhD student), Thierry Arrabal, Olivier Simonin, Stephane d'Alu, Isabelle Guerin-Lassous (Lyon 1/Lip), Nicolas Valle, Hervé Rivano (Agora team).

In the context of the last year of the CONCERTO ANR project, we studied the impact of the quality of Wi-Fi communications on the behavior of remotely controlled UAVs. We designed an experimental platform composed of PX4 Vision UAVs, a Motion Capture system (the one of the ARMEN/LS2N team), and a tool we developed to generate mission traffic from the ground station. This allowed us to analyze how the presence of mission traffic impacts the stability of the reception of the control traffic, and how, in turn, it impacts the behavior of the UAVs. This work is presented in the publication 20 and illustrated in Fig. 10.

Show image description

Traffic perturbation on UAV remote control

Show image description

Traffic perturbation on UAV remote control

In the context of the PhD thesis of Théotime Balaguer, we proposed in 2023 a co-simulator, called DANCERS, allowing to connect any physics robotic simulator with any network simulator. In 2024 we finalized the implementation and its evaluation, in particular by connecting Gazebo and ns3 simualtors to study communication in multi-robot systems such as in a fleet of drones. DANCERS allows also SITL hybrid simulation (Software in the loop). This work has been published in the french revue ROIA 14 and recently accepted for publication in the international conference IEEE SIMPAR 2025 (Simulation, Modeling, and Programming for Autonomous Robots).

In 2024, we continued to study the Time of Flight approach to measure distances between UAVs. Stephane D'Alu, who is preparing a PhD in both Chroma and Agora teams, study the implementation of a broadcast-based protocol 48 with UWB Decawave chips embedded on UAVs. The proposed solution has been simulated and also experimented with several static units. A first publication is in prepapration.

8.6.2 Flocking-based UAVs chain creation

Participants: Théotime Balaguer (PhD student), Olivier Simonin, Isabelle Fantoni (LS2N, Nantes), Isabelle Guerin-Lassous (Lyon 1/Lip), Etienne Villemurev(Univ. Sherbrooke/3IT, Canada).

PhD thesis of Theotime Balaguer

Show image description

Chain creation

Show image description

Chain creation

In the context of the PhD thesis of Théotime Balaguer and the Associated Team ROBLOLC with Univ. of Sherbrooke, we proposed a new decentralized algorithm to manage the deployment of a fleet of drones from a source area to one or several target areas. Our approach relies on the extension of the flocking model, which was initially defined to allow a fleet of agents to fly together. We propose an algorithm ensuring the emergence of a “mission” communication path from a leader drone to a source drone while the leader flies towards its target area. By defining two possible roles for others drones (mission node/free node), drones dynamically integrates the “mission” path while the flocking stretches while avoiding obstacles and their communication perturbations. The originality of our model is to be decentralized and able to exploit data routing measures to optimize dynamically the mission paths in term of distance and QoS. This work is illustrated in figure 11. A publication is in preparation.

8.6.3 Graph-based decentralized exploration with a swarm of UAVs in unknown subterrean environments

Participants: Mathis Fleuriel (PhD student), Olivier Simonin, Elena Vanneaux (ENSTA Paris, U2IS), David Filiat (ENSTA Paris, U2IS).

The use of unmanned aerial systems (UAS) is a promising solution for search and rescue (SAR) operations conducted to find missing persons. In the ANR VORTEX project, we study how it is possible to deploy rapidly autonomous drones in indoor environments without using conventional communications or mapping but by leveraging cameras & vision-based processes. We propose to explore an approach where a swarm of small UAVs (quadrotors) will quickly deploy in the environment while forming a dynamic mesh of sensors. By flying autonomously but keeping visual contact with at least another UAV, each drone will be able to estimate its relative location but also to communicate information from drone to drone. As a consequence, drones will form a dynamic topological graph covering progressively the environment.

In 2024, we started to study epxloration stategies through Mathis Fleuriel's Master interbnship (co-supervized by E. Vanneaux from ENSTA Paris and Olivier Simonin from Chroma team). We proposed a first model, defined as multi-agent behaviors allowing the swarm to explore the environment and to self-reconfigure when perturbed (eg. when a dead-end is encountred). Our approach is to merge graph search strategies and swarm behaviors to achieve time efficiency while managing the limited number of drones. To this end, the model relies on deploying a chain of drones guided by a leader which is the drone managing the exploration (the leader role can change when the chain needs reconfiguration). In the end of 2024, Mathis Fleuriel was recruited as PhD student, in Lyon, in the VORTEX project, to continue this research (with same co-supervisors). A first publication of the chain model is in preparation.

8.6.4 Robotic Matter

Participants: Baudouin Saintyves, Heinrich Jaeger (Prof. University of Chicago), Olivier Simonin.

Show image description

Tensile test experiment on a passive proof of concept, toward the conception of Shellbots. Full deformation cycle from left to right.

Since November, Baudouin Saintyves made progress on the development of a new robotic prototype, the Shellbot. The goal is to build a monolayer that can reconfigure robotically to change its local curvature. We have built a passive prototype, a metamaterial, and performed preliminary mechanical testing to help the design of a future robotic version (see figure 12). This study consists in exploring how the reorganization of the building blocks is correlated with the final curvature of the material.

8.7.1 Optimally Solving Decentralized POMDPs Under Hierarchical Information Sharing

Participants: Johan Peralez, Aurélien Delage, Olivier Buffet (Inria Nancy, Larsen), Jilles S. Dibangoye.

This work is the result of an effort to provide a strong theory for optimally solving decentralized decision-making problems, represented as decentralized partially observable Markov decision processes (Dec-POMDPs), with tractable structures 65. This work describes core ideas of multi-agent reinforcement learning algorithms for decentralized partially observable Markov decision processes (Dec-POMDPs) under hierarchical information sharing. That is the dominant management style in our society—each agent knows what her immediate subordinate knows. In this case, we prove backups to be linear due to (normal-form) Bellman optimality equations being for the first time specified in extensive form. Doing so, one can sequence agent decisions at every point in time while still preserving optimality; hence paving the way for the first optimal $Q$-learning algorithm with linear-time updates for an important subclass of Dec-POMDPs. That is a significant improvement over the double-exponential time required for normal-form Bellman optimality equations, in which updates of agent decisions are performed in sync every point in time. Experiments show algorithms utilizing extensive instead of normal-form Bellman optimality equations can be applied effectively to much larger problems. This work has been published in ICML 2024 28.

8.7.2 Optimally Solving Decentralized POMDPs: The Sequential Central Learning Approach

Participants: Johan Peralez [Post-doc], Aurélien Delage [PhD Student], Olivier Buffet [Inria Nancy, Larsen], Jilles S. Dibangoye.

The centralized learning for simultaneous decentralized execution paradigm emerged as the state-of-the-art approach to $ϵ$-optimally solving a simultaneous- move decentralized partially observable Markov decision process, but the scalability remains a significant issue. This paper presents a novel—yet more scalable—alternative, namely sequential central learning for simultaneous decentralized execution. This methodology pushes the applicability of Bellman’s principle of optimality further, raising three new properties. First, it allows a central learner to reason upon sufficient sequential statistics instead of prior simultaneous ones. Next, it proves that $ϵ$-optimal value functions are piecewise linear and convex in sufficient sequential statistics. Finally, it drops the time complexity of the backup operators from double exponential to linear at the expense of longer planning horizons. Besides, it makes it easy to use single-agent methods—e.g., a reinforcement learning algorithm enhanced with these findings applies while still preserving its asymptotic convergence guarantees. Experiments on standard 2- as well as $n$-agent benchmarks from the literature against state-of-the-art $ϵ$-optimal solvers confirm the superiority of the novel paradigm. This work was accepted for publication in AAAI 2025 conference.

8.8.1 Constraint Programming

Participants: Jean-Yves Courtonne (Inria STEEP), François Delobel (LIMOS), Patrick Derbez (Inria Capsule), Léon Fauste, Arthur Gontier (Inria Capsule), Mathieu Mangeot (Inria STEEP), Loïc Rouquette, Christine Solnon.

In many of our applications, we have to solve constrained optimization problems (COPs) such as vehicle routing problems, or multi-agent path finding problems, for example. These COPs are usually NP-hard, and we study and design algorithms for solving these problems. We more particularly study approaches based on Constraint Programming (CP), that provides declarative languages for describing these problems by means of constraints, and generic algorithms for solving them by means of constraint propagation. In 61, we introduce canonical codes for describing benzenoids, and we introduce CP models for efficiently generating all benzenoids that satisfy some given constraints.

PhD thesis of Léon Fauste (since Sept. 2021). This thesis is funded by Ecole Normale Supérieure Paris Saclay and it is co-supervised by Mathieu Mangeot, Jean-Yves Courtonne (from Inria STEEP team) and Christine Solnon. The goal of the thesis is to design new tools for choosing the best geographical scale when relocating productive activities, and the context of this work is described in 49. During the first months of his thesis, Léon has designed a first Integer Linear Programming (ILP) model described in 50. In 2023/2024, Léon is on leave (césure) for one year in order to follow a Master in Geography.

8.8.2 Vehicle routing problems

Participants: Romain Fontaine (PhD Student CITI), Christine Solnon, Jilles S. Dibangoye (Univ. Groningen).

The research is done within the context of the transportation challenge and funded by INSA Lyon.

9.1.1 IRT Nanoelec PULSE – Security of Autonomous Mobile Robotics (2023-2025)

Participants: Lukas Rummelhard, Thomas Genevois, Robin Baruffa, Hugo Bantignies, Nicolas Turro, Christian Laugier.

Chroma participates in a project supported by PIA in the scope of the program PULSE of IRT Nanoelec. The objective of this project is to integrate, develop and promote technological bricks of context capture, for the safety of the autonomous vehicle. Building on Embedded Bayesian Perception for Dynamic Environment, Bayesian data fusion and filtering technologies from sets of heterogeneous sensors, these bricks make it possible to secure the movements of vehicles, but also provide them with an enriched and useful representation for autonomy functions themselves. In this context, various demonstrators embedding those technology bricks are developed in cooperation with industrial partners. Current work is mainly focused on distributed perception using realistic communication systems and Bayesian occupancy grids, and integration of Neural-Network-based systems into probabilistic frameworks.

9.1.2 Start-up maturation : Yona-Robotics (2024-2025)

Participants: Lukas Rummelhard, Christian Laugier.

In order to address the emerging markets for mobile robotics, in particular in industrial contexts (industrial robotics, handling, logistics, agricultural robotics, inspection robots, service robots, etc.), a technology maturation project on the Embedded Bayesian Perception and Navigation systems of the team (CMCDOT framework) have started this year, thanks to the support of the SATT LINKSIUM GRENOBLE ALPES. See Yona-Robotics.

10.1.1 Inria associate team not involved in an IIL or an international program

10.2.1 Visits of international scientists

10.2.2 Visits to international teams

10.2.3 H2020 projects

10.3.1 ANR

10.3.2 PEPR

10.3.3 Grand Defis

10.3.4 INRIA/AID AI projects

11.1.1 Scientific events: selection

11.1.2 Journal

11.1.3 Invited talks

• Alessandro Renzaglia gave an invited talk titled "Robust and adaptive solutions for multi-robot coordination" at the IROS 2024 Workshop on Real-World Challenges in Multi-Robot Cooperation.

11.1.4 Leadership within the scientific community

• Olivier Simonin has been appointed as member of the scientific board of the GDR Robotique (starting 1/1/2025).
• Anne Spalanzani chaired the ANR CE33 Robotics and Interaction committee.
• Anne Spalanzani co-animates the theme TS5 "Robotique centrée sur les données" in GDR Robotique.
• Christian Laugier is member of the GDR Robotique "comité des sages".
• Olivier Simonin is member of the Executive Committee of the PEPR O2R (Organic Robotics) 2023-2030 (www.cnrs.fr/en/pepr/pepr-exploratoire-o2r-robotique).
• Fabrice Jumel is a member of the organization committee of the international Robocup@Home competition (since 2016).

11.1.5 Scientific expertise

• Anne Spalanzani is a jury member of the Georges Giralt best European PhD in Robotic.
• Olivier Simonin was member of the HCERES committee for the evalation of the CRISTAL Lab. (Lille).

11.1.6 Research administration

• O. Simonin is member of the Inria Lyon COS (comité d'orientation scientifique du Centre).
• O. Simonin is the PhD student representative for the CITI laboratory at the InfoMath doctoral school in Lyon.
• Anne Spalanzani is elected to the "pôle de recherche MSTIC" of Grenoble-Alpes University (2025-2029).
• Anne Spalanzani is member of the mentoring committee of the Grenoble-Alpes University (mentoring 1 or 2 collegues per year).

11.2.1 Supervision

Phd in progress :

• Mathis Fleuriel,"Exploration coopérative et distribuée en intérieur par des flottes de drones avec des communications limitées", co-supervized by Olivier Simonin (Chroma) and Elena Vanneaux (ENSTA Paris, U2IS), ANR "VORTEX" grant.
• Simon Ferrier, "Planification d'itinéraires multi-robots hétérogènes pour des opérations culturales", co-supervized by Olivier Simonin and Alessandro Renzaglia, ANR NINSAR/PEPR "AgroEcol. & Numerique" grant.
• Kaushik Bhowmik, "Modeling and prediction of pedestrian behavior on bikes, scooters or hoverboards" started on 1/05/2023. Co-supervized by A. Spalanzani (Chroma) and P. Martinet (Acentauri, Sophia Antipolis).
• Benjamin Sportich, Next-Best-View Planning for 3D Reconstruction with Cooperative Multi-UAV Systems, A. Renzaglia, O. Simonin (ANR AVENUE).
• Théotime Balaguer, Contrôle efficace de flotte de drones soumis à des contraintes de communication hétérogènes, O. Simonin, I. Guérin Lassous, I. Fantoni (Lyon1/LIP), (CNRS/LS2N Nantes) (INSA/DGA).
• Rabbia Asghar, Real-time analysis of complex traffic situations by Predicting/Characterizing Abnormal Behaviors and Dangerous Situations, A. Spalanzani, C. Laugier, L. Rummerhard.
• Manuel Diaz Zapata, AI-based Framework for Scene Understanding and Situation Awareness for Autonomous Vehicles, J. Dibangoye, C. Laugier, O. Simonin.
• Jean-Baptiste Horel, Validation des composants de perception basés sur l’IA dans les véhicules autonomes, R. Mateescu (Inria Convecs), A. Renzaglia, C. Laugier (funded by PRISSMA project).
• Léon Fauste, Aide à la décision pour le choix de l'échelle lors de la relocalisation d'activités industrielles, M. Mangeot (STEEP, Inria Montbonnot), J.-Y. Courtonne (STEEP, Inria Montbonnot), and C. Solnon.
• Gustavo Salazar Gomez, AI-driven safe motion planning and driving decision-making for autonomous driving. Thèse de l'Université Grenoble Alpes. A. Spalanzani, C. Laugier.

Phd defended in 2024 :

• Xiao Peng, "Planning Problems in a Multi-Tethered-Robot System", O. Simonin, C. Solnon, European H2020 BugWright2 project funding. Defended on 19/2/24.
• Thomas Genevois, "Evitement de collisions en robotique, basé sur grilles d'occupation dynamiques". Thèse en validation d'acquis, Université Grenoble Alpes. A. Spalanzani, C. Laugier. Defended on 19/06/24.
• Pierre Marza, "Learning spatial representations for single-task navigation and multi-task policies", C. Wolf (NaverLabs Europe), L. Matignon (Lyon1/LIRIS), O. Simonin (funded by Chaire IA "REMEMBER"). Defended on 25/11/24.
• Aurelien Delage, "Theoretical foundations of planning in partially observable stochastic games", Jilles Dibangoye (Chroma) and Olivier Buffet (Larsen Inria Nancy), funded by ANR JCJC PLASMA. Defended on 28/06/24.
• Romain Fontaine, "Exact and anytime heuristic search for the Time Dependent Traveling Salesman Problem with Time Windows", J. Dibangoye and C. Solnon. Defended on 9/07/2024.

11.2.2 Juries

Anne Spalanzani reported on the following PhD and HDR:

• Mehdi Nima, Thèse de l'Université de Nancy, Décembre 2024
• Mahsa Golchoubian, thèse de l'Université de Waterloo, Canada, Novembre 2024
• Benjamin Camblor, Thèse de l'Université de Bordeaux, avril 2024.
• Sylvain Guégan, Habilitation à diriger des recherches, INSA Rennes Septembre 2024.

Anne Spalanzani examined the following PhD

• Imane Argui, Thèse de l'Université de Rouen, Décembre 2024, présidente du jury
• Anand Ballou, Thèse de l'Université Grenoble Alpes, Mars 2024, présidente du jury

Olivier Simonin reported on the following PhD and HDR:

• Fabio Morbidi, Habilitation à diriger des recherches de l'Université de Picardie Jules Verne, Amiens, January 31th, 2024.
• Mohammed Rahmani, Thèse de l'Université de Bourgogne, Nevers, October 21th, 2024.

Olivier Simonin examined the following PhD

• Maxime Toquebiau, Thèse de Sorbonne Université, December 13th, 2024.
• Abbass Chreim, Thèse de l'Université de Lille, December 5th, 2024.
• Déborah Conforto Nedelmann, Thèse de l'Université de Toulouse, ISAE-SUPAERO, November 26th, 2024.
• Joris Dinneweth, Thèse de l'Université Paris-Saclay, December 9th, 2024.
• Lyes Saidi, Thèse de l'UTC, Compiegne, January 10th, 2024.
• Matěj Petrlik, Thèse de l'Université de Prague/CTU, Rep. Tchèque. December 11th, 2024.

International journals

• 14 articleT.Théotime Balaguer, O.Olivier Simonin, I.Isabelle Guérin-Lassous and I.Isabelle Fantoni. Etat de l'art sur la co-simulation robotique et réseau des systèmes multi-robots.Revue Ouverte d'Intelligence Artificielle2024, 1-25In press. HALback to text
• 15 articleA.Agostino Martinelli. Corrigendum to "Nonlinear unknown input observability and unknown input reconstruction: The general analytical solution".Information FusionOctober 2024, 1-3HALDOIback to text
• 16 articleA.Agostino Martinelli. Revisiting the observability and identifiability properties of a popular HIV model.Journal of Theoretical BiologyJune 2024, 1-29HALDOIback to text
• 17 articleX.Xiao Peng, F.François Schwarzentruber, O.Olivier Simonin and C.Christine Solnon. Spanning-tree based coverage for a tethered robot.IEEE Robotics and Automation LettersDecember 2024, 1-8In press. HALback to text
• 18 articleL.Loïc Rouquette, M.Marine Minier and C.Christine Solnon. Automatic boomerang attacks search on Rijndael.Journal of Mathematical Cryptology181February 2024, 1-16HALDOI
• 19 articleP. T.Phani Teja Singamaneni, P.Pilar Bachiller-Burgos, L.Luis Manso, A.Anaís Garrell, A.Alberto Sanfeliu, A.Anne Spalanzani and R.Rachid Alami. A Survey on Socially Aware Robot Navigation: Taxonomy and Future Challenges.The International Journal of Robotics Research2024, 1-33In press. HALDOI

International peer-reviewed conferences

• 20 inproceedingsT.Thierry Arrabal, T.Théotime Balaguer, I.Isabelle Guérin-Lassous and O.Olivier Simonin. Experimental analysis of Wi-Fi-based remote control of UAVs with concurrent mission traffic.International conference on distributed computing in smart systems and the internet of thingsWi-DroIT 2024 - Wireless Sensors and Drones in the Internet of ThingsAbu Dhabi, United Arab Emirates2024, 1-8HALback to text
• 21 inproceedingsR.Rabbia Asghar, W.Wenqian Liu, L.Lukas Rummelhard, A.Anne Spalanzani and C.Christian Laugier. Flow-guided Motion Prediction with Semantics and Dynamic Occupancy Grid Maps.ITSC 2024 - 27th IEEE International Conference on Intelligent Transportation SystemsEdmonton, CanadaIEEE2024, 1-6HALDOIback to text
• 22 inproceedingsL.Lara Briñón-Arranz, M.Martin Abou Hamad and A.Alessandro Renzaglia. Voronoi-based Multi-Robot Formations for 3D Source Seeking via Cooperative Gradient Estimation.ICARCV 2024 - 18th International Conference on Control, Automation, Robotics and VisionDubai, United Arab Emirates2024, 1-6HALback to text
• 23 inproceedingsM. A.Manuel Alejandro Diaz-Zapata, W.Wenqian Liu, R.Robin Baruffa and C.Christian Laugier. BEVal: A Cross-dataset Evaluation Study of BEV Segmentation Models for Autonomous Driving.ICARCV 2024 - 18th International Conference on Control, Automation, Robotics and Vision - ICARCV 2024Dubai, United Arab Emirates2024, 1-3HALback to text
• 24 inproceedingsE.Erwan Escudie, L.Laëtitia Matignon and J.Jacques Saraydaryan. Attention Graph for Multi-Robot Social Navigation with Deep Reinforcement Learning.Proceedings of the 23rd International Conference on Autonomous Agents and Multiagent SystemsAAMAS 2024 - International Conference on Autonomous Agents and Multi-Agent SystemsAuckland, New Zealand2024HALback to text
• 25 inproceedingsJ.-B.Jean-Baptiste Horel, A.Alessandro Renzaglia, R.Radu Mateescu and C.Christian Laugier. Scenario-based Validation of Autonomous Vehicles using Augmented Reality.ICRA 2024 - 40th Anniversary of the IEEE Conference on Robotics and AutomationRotterdam, Netherlands2024, 1-3HALback to text
• 26 inproceedingsP.Pierre Marza, L.Laetitia Matignon, O.Olivier Simonin, D.Dhruv Batra, C.Christian Wolf and D. S.Devendra Singh Chaplot. AutoNeRF: Training Implicit Scene Representations with Autonomous Agents.IROS 2024 - International Conference on Intelligent Robots and SystemsAbu Dhabi, United Arab Emirates2024, 1-8HALback to text
• 27 inproceedingsP.Pierre Marza, L.Laetitia Matignon, O.Olivier Simonin and C.Christian Wolf. Task-conditioned adaptation of visual features in multi-task policy learning.CVPR 2024 - IEEE/CVF Conference on Computer Vision and Pattern RecognitionSeattle, United StatesIEEE2024, 1-16HALback to text
• 28 inproceedingsJ.Johan Peralez, A.Aurélien Delage, O.Olivier Buffet and J.Jilles Dibangoye. Solving Hierarchical Information-Sharing Dec-POMDPs: An Extensive-Form Game Approach.Proceedings of the 41st International Conference on Machine LearningICML 2024 - 41st International Conference on Machine Learning235Proceedings of Machine Learning ResearchVienne, Austria2024, 63010HALback to text
• 29 inproceedingsB.Benoit Renault, J.Jacques Saraydaryan, D.David Brown and O.Olivier Simonin. Multi-Robot Navigation among Movable Obstacles: Implicit Coordination to Deal with Conflicts and Deadlocks.IROS 2024 - IEEE/RSJ International Conference on Intelligent Robots and SystemsAbu DHABI, United Arab EmiratesIEEE2024, 1-7HALback to text
• 30 inproceedingsG.Gustavo Salazar-Gomez, W.Wenqian Liu, M. A.Manuel Alejandro Diaz-Zapata, D.David Sierra González and C.Christian Laugier. TLCFuse: Temporal Multi-Modality Fusion Towards Occlusion-Aware Semantic Segmentation.IV 2024 - 35th IEEE Intelligent Vehicles SymposiumJeju Island, South KoreaIEEE2024, 2110-2116HALDOIback to text
• 31 inproceedingsJ.Jacques Saraydaryan, F.Fabrice Jumel and O.Olivier Simonin. Prédiction des flux de personnes par propagation sur GPU des observations d'une flotte de robots.JFSMA 2024JFSMA 2024 - 32èmes Journées Francophones sur les Systèmes Multi-AgentsCargèse, France2024, 1-10HALback to text

Doctoral dissertations and habilitation theses

• 32 thesisA.Aurélien Delage. Theoretical foundations for planning in partially observable stochastic games.INSA de LyonJune 2024HAL
• 33 thesisR.Romain Fontaine. Exact and anytime heuristic search for the Time Dependent Traveling Salesman Problem with Time Windows.INSA de LyonJuly 2024HALback to text
• 34 thesisT.Thomas Genevois. Dynamic Occupancy Grid based Collision Avoidance in Robotics.Université Grenoble Alpes [2020-....]June 2024HAL
• 35 thesisP.Pierre Marza. Learning spatial representations for single-task navigation and multi-task policies.INSA de LyonNovember 2024HALback to textback to text
• 36 thesisX.Xiao Peng. Planning problems in a multi-tethered-robot system.INSA de LyonFebruary 2024HALback to text

Reports & preprints

• 37 miscA.Agostino Martinelli. Detection of a serious error in the paper: "On identifiability of nonlinear ODE models and applications in viral dynamics".January 2024HAL
• 38 miscA.Agostino Martinelli. General analytical condition to nonlinear identifiability and its application in viral dynamics.January 2024HAL
• 39 miscA.Agostino Martinelli. Identifiability of nonlinear ODE Models with Time-Varying Parameters: the General Analytical Solution and Applications in Viral Dynamics.January 2024HAL
• 40 miscA.Agostino Martinelli. Improving Unknown Input canonization in UIO.2024HALback to text

Other scientific publications

• 41 inproceedingsR.Rabbia Asghar, W.Wenqian Liu, L.Lukas Rummelhard, A.Anne Spalanzani and C.Christian Laugier. Integrating Specialized and Generic Agent Motion Prediction with Dynamic Occupancy Grid Maps.IROS 2024 - 37th IEEE/RSJ International Conference on Intelligent Robots and SystemsAbu Dhabi, United Arab EmiratesIEEE2024, 1-1HALback to text
• 42 miscJ.-B.Jean-Baptiste Horel, A.Alessandro Renzaglia, R.Radu Mateescu and C.Christian Laugier. Scenario-based Validation of Autonomous Vehicles using Augmented Reality: Video of the Experiments.March 2024HAL
• 43 inproceedingsD.Damien Rambeau, C.Chadi M'Sila, S.Simon Ferrier, O.Olivier Simonin and C.Cédric Pradalier. Hybrid control framework for tethered multi-robot systems: Poster/Video 4min.ICRA 2024 – 40th IEEE Conference on Robotics and AutomationRotterdam, NetherlandsIEEE2024, 1-1HALback to text

Scientific popularization

• 44 articleC.Christine Solnon. Les progrès technologiques créent une augmentation des besoins.La Recherche578July 2024, 1-2HAL
• 45 article C.Christine Solnon. Optimiser, mais pour qui ? La Recherche January 2024 HAL
• 46 articleC.Christine Solnon. Reféminiser le monde de la programmation.La Recherche5772024HAL