Laura Ferranti

Andrei-Carlo Papuc, Lasse Peters, Sihao Sun et al.

Autonomous drone racing pushes the boundaries of high-speed motion planning and multi-agent strategic decision-making. Success in this domain requires drones not only to navigate at their limits but also to anticipate and counteract competitors' actions. In this paper, we study a fundamental question that arises in this domain: how deeply should an agent strategize before taking an action? To this end, we compare two planning paradigms: the Model Predictive Game (MPG), which finds interaction-aware strategies at the expense of longer computation times, and contouring Model Predictive Control (MPC), which computes strategies rapidly but does not reason about interactions. We perform extensive experiments to study this trade-off, revealing that MPG outperforms MPC at moderate velocities but loses its advantage at higher speeds due to latency. To address this shortcoming, we propose a Learned Model Predictive Game (LMPG) approach that amortizes model predictive gameplay to reduce latency. In both simulation and hardware experiments, we benchmark our approach against MPG and MPC in head-to-head races, finding that LMPG outperforms both baselines.

Nikhil Sethi, Max Lodel, Laura Ferranti et al.

Autonomous target search is crucial for deploying Micro Aerial Vehicles (MAVs) in emergency response and rescue missions. Existing approaches either focus on 2D semantic navigation in structured environments -- which is less effective in complex 3D settings, or on robotic exploration in cluttered spaces -- which often lacks the semantic reasoning needed for efficient target search. This paper overcomes these limitations by proposing a novel framework that utilizes a semantically-guided viewpoint planner to minimize target search and exploration time in unstructured 3D environments using an MAV. Specifically, we develop a combinatorial planner that generates efficient semantic exploration plans by prioritizing viewpoints that likely lead to the target. To guide the planner towards the target, an active perception pipeline is developed that propagates semantic priorities of observed objects into neighboring frontier voxels for computing semantic information gains of frontier viewpoints. In addition, we demonstrate how LLM-based similarity scores can be leveraged as semantic priority input to our pipeline. Evaluations in two distinct simulation environments show that the proposed method consistently outperforms baselines by quickly finding the target while maintaining reasonable exploration times. Real-world experiments with an MAV further demonstrate the method's ability to handle practical constraints like limited battery life, small sensor range, and semantic uncertainty.

Max Lodel, Bruno Brito, Álvaro Serra-Gómez et al.

Search missions require motion planning and navigation methods for information gathering that continuously replan based on new observations of the robot's surroundings. Current methods for information gathering, such as Monte Carlo Tree Search, are capable of reasoning over long horizons, but they are computationally expensive. An alternative for fast online execution is to train, offline, an information gathering policy, which indirectly reasons about the information value of new observations. However, these policies lack safety guarantees and do not account for the robot dynamics. To overcome these limitations we train an information-aware policy via deep reinforcement learning, that guides a receding-horizon trajectory optimization planner. In particular, the policy continuously recommends a reference viewpoint to the local planner, such that the resulting dynamically feasible and collision-free trajectories lead to observations that maximize the information gain and reduce the uncertainty about the environment. In simulation tests in previously unseen environments, our method consistently outperforms greedy next-best-view policies and achieves competitive performance compared to Monte Carlo Tree Search, in terms of information gains and coverage time, with a reduction in execution time by three orders of magnitude.

Bas van der Heijden, Jelle Luijkx, Laura Ferranti et al.

Sim2real, that is, the transfer of learned control policies from simulation to real world, is an area of growing interest in robotics due to its potential to efficiently handle complex tasks. The sim2real approach faces challenges due to mismatches between simulation and reality. These discrepancies arise from inaccuracies in modeling physical phenomena and asynchronous control, among other factors. To this end, we introduce EAGERx, a framework with a unified software pipeline for both real and simulated robot learning. It can support various simulators and aids in integrating state, action and time-scale abstractions to facilitate learning. EAGERx's integrated delay simulation, domain randomization features, and proposed synchronization algorithm contribute to narrowing the sim2real gap. We demonstrate (in the context of robot learning and beyond) the efficacy of EAGERx in accommodating diverse robotic systems and maintaining consistent simulation behavior. EAGERx is open source and its code is available at https://eagerx.readthedocs.io.

Jelle Luijkx, Zlatan Ajanovic, Laura Ferranti et al.

Several recent works show impressive results in mapping language-based human commands and image scene observations to direct robot executable policies (e.g., pick and place poses). However, these approaches do not consider the uncertainty of the trained policy and simply always execute actions suggested by the current policy as the most probable ones. This makes them vulnerable to domain shift and inefficient in the number of required demonstrations. We extend previous works and present the PARTNR algorithm that can detect ambiguities in the trained policy by analyzing multiple modalities in the pick and place poses using topological analysis. PARTNR employs an adaptive, sensitivity-based, gating function that decides if additional user demonstrations are required. User demonstrations are aggregated to the dataset and used for subsequent training. In this way, the policy can adapt promptly to domain shift and it can minimize the number of required demonstrations for a well-trained policy. The adaptive threshold enables to achieve the user-acceptable level of ambiguity to execute the policy autonomously and in turn, increase the trustworthiness of our system. We demonstrate the performance of PARTNR in a table-top pick and place task.

Mohammed Irshadh Ismaaeel Sathyamangalam Imran, Lasse Peters, Michael Khayyat et al.

We address the multi-agent motion planning problem where interactions, collisions, and congestion co-exist. Conventional game-theoretic planners capture interactions among agents but often converge to conservative, congested equilibria. Homotopy planners, on the other hand, can explore topologically distinct paths, but lack mechanisms to account for the interdependence of agents' future actions. We propose a unified framework that leverages homotopy classes as structured strategy sets within a receding-horizon setup. At each planning stage, a deterministic homotopy planner generates topologically distinct paths for each agent, conditioned on the joint configuration. To avoid intractable growth of candidate paths, we propose a simple heuristic filtering step that selects a top-$K$ subset of the most suitable congestion-free joint strategies to ensure computational tractability. These serve as initializations for a potential game that enforces homotopy-consistent constraints and yields a generalized open-loop Nash equilibrium (OLNE), with penalties discouraging abrupt strategy shifts in a receding-horizon setting. Simulations with three agents demonstrate improved efficiency (faster completion) and enhanced safety (greater inter-agent clearance, leading to reduced congestion) compared to a local baseline and NH-ORCA that do not reason about homotopies. Hardware trials with two robots and one human demonstrate robustness to irrational behaviors, where our method adapts by switching to alternative feasible equilibria while the baseline game fails.

Lasse Peters, Laura Ferranti, Javier Alonso-Mora et al.

Imitation learning powered by generative models has proven effective for modeling complex single-agent behaviors. However, teaching multi-agent systems, like multiple arms or vehicles, to coordinate through imitation learning is hindered by a fundamental data bottleneck: as the joint state-action space grows exponentially with the number of agents, collecting a sufficient amount of coordinated multi-agent demonstrations becomes extremely costly. In this work, we ask: how can we leverage single-agent demonstration data to learn multi-agent policies? We present Coordinated Diffusion (CoDi), a framework that couples independently trained single-agent diffusion policies through a user-defined multi-agent cost function, without requiring any coordinated demonstrations. We derive a new diffusion-based sampling scheme wherein the diffusion score function decomposes into independent, single-agent pre-trained base policies plus a cost-driven guidance term that coordinates these base policies into cohesive multi-agent behavior. We show that this guidance term can be estimated in a gradient-free manner, making CoDi applicable to black-box, non-differentiable cost functions without additional training. Theoretically and empirically, we analyze the conditions under which this composition can faithfully approximate a target multi-agent behavior. We find a complementary role for demonstration data versus the cost function: single-agent demonstrations must cover the support of the desired multi-agent behavior, while the cost function must promote desired behavior from this product of single-agent policies. Our results in simulation and hardware experiments of a two-arm manipulation task show that CoDi discovers robust coordinated behavior from single-agent data, is more data-efficient than multi-agent baselines, and highlights the importance of joint guidance, base policy support, and cost design.

Jelle Luijkx, Zlatan Ajanović, Laura Ferranti et al.

Human teaching effort is a significant bottleneck for the broader applicability of interactive imitation learning. To reduce the number of required queries, existing methods employ active learning to query the human teacher only in uncertain, risky, or novel situations. However, during these queries, the novice's planned actions are not utilized despite containing valuable information, such as the novice's capabilities, as well as corresponding uncertainty levels. To this end, we allow the novice to say: "I plan to do this, but I am uncertain." We introduce the Active Skill-level Data Aggregation (ASkDAgger) framework, which leverages teacher feedback on the novice plan in three key ways: (1) S-Aware Gating (SAG): Adjusts the gating threshold to track sensitivity, specificity, or a minimum success rate; (2) Foresight Interactive Experience Replay (FIER), which recasts valid and relabeled novice action plans into demonstrations; and (3) Prioritized Interactive Experience Replay (PIER), which prioritizes replay based on uncertainty, novice success, and demonstration age. Together, these components balance query frequency with failure incidence, reduce the number of required demonstration annotations, improve generalization, and speed up adaptation to changing domains. We validate the effectiveness of ASkDAgger through language-conditioned manipulation tasks in both simulation and real-world environments. Code, data, and videos are available at https://askdagger.github.io.

Smit Saparia, Andreas Schimpe, Laura Ferranti

Autonomous cars can reduce road traffic accidents and provide a safer mode of transport. However, key technical challenges, such as safe navigation in complex urban environments, need to be addressed before deploying these vehicles on the market. Teleoperation can help smooth the transition from human operated to fully autonomous vehicles since it still has human in the loop providing the scope of fallback on driver. This paper presents an Active Safety System (ASS) approach for teleoperated driving. The proposed approach helps the operator ensure the safety of the vehicle in complex environments, that is, avoid collisions with static or dynamic obstacles. Our ASS relies on a model predictive control (MPC) formulation to control both the lateral and longitudinal dynamics of the vehicle. By exploiting the ability of the MPC framework to deal with constraints, our ASS restricts the controller's authority to intervene for lateral correction of the human operator's commands, avoiding counter-intuitive driving experience for the human operator. Further, we design a visual feedback to enhance the operator's trust over the ASS. In addition, we propose an MPC's prediction horizon data based novel predictive display to mitigate the effects of large latency in the teleoperation system. We tested the performance of the proposed approach on a high-fidelity vehicle simulator in the presence of dynamic obstacles and latency.

Bas van der Heijden, Laura Ferranti, Jens Kober et al.

This paper presents DeepKoCo, a novel model-based agent that learns a latent Koopman representation from images. This representation allows DeepKoCo to plan efficiently using linear control methods, such as linear model predictive control. Compared to traditional agents, DeepKoCo learns task-relevant dynamics, thanks to the use of a tailored lossy autoencoder network that allows DeepKoCo to learn latent dynamics that reconstruct and predict only observed costs, rather than all observed dynamics. As our results show, DeepKoCo achieves similar final performance as traditional model-free methods on complex control tasks while being considerably more robust to distractor dynamics, making the proposed agent more amenable for real-life applications.

Bruno Brito, Boaz Floor, Laura Ferranti et al.

This paper presents a method for local motion planning in unstructured environments with static and moving obstacles, such as humans. Given a reference path and speed, our optimization-based receding-horizon approach computes a local trajectory that minimizes the tracking error while avoiding obstacles. We build on nonlinear model-predictive contouring control (MPCC) and extend it to incorporate a static map by computing, online, a set of convex regions in free space. We model moving obstacles as ellipsoids and provide a correct bound to approximate the collision region, given by the Minkowsky sum of an ellipse and a circle. Our framework is agnostic to the robot model. We present experimental results with a mobile robot navigating in indoor environments populated with humans. Our method is executed fully onboard without the need of external support and can be applied to other robot morphologies such as autonomous cars.

Roya Firoozi, Laura Ferranti, Xiaojing Zhang et al.

This work presents a distributed method for multi-vehicle coordination based on nonlinear model predictive control (NMPC) and dual decomposition. Our approach allows the vehicles to coordinate in tight spaces (e.g., busy highway lanes or parking lots) by using a polytopic description of each vehicle's shape and formulating collision avoidance as a dual optimization problem. Our method accommodates heterogeneous teams of vehicles (i.e., vehicles with different polytopic shapes and dynamic models can be part of the same team). Our method allows the vehicles to share their intentions in a distributed fashion without relying on a central coordinator and efficiently provides collision-free trajectories for the vehicles. In addition, our method decouples the individual-vehicles' trajectory optimization from their collision-avoidance objectives enhancing the scalability of the method and allowing one to exploit parallel hardware architectures. All these features are particularly important for vehicular applications, where the systems operate at high-frequency rates in dynamic environments. To validate our method, we apply it in a vehicular application, that is, the autonomous lane-merging of a team of connected vehicles to form a platoon. We compare our design with the centralized NMPC design to show the computational benefits of the proposed distributed algorithm.