Zlatan Ajanović

Exploration in environments with sparse rewards remains a fundamental challenge in reinforcement learning (RL). Existing approaches such as curriculum learning and Go-Explore often rely on hand-crafted heuristics, while curiosity-driven methods risk converging to suboptimal policies. We propose Search-Inspired Exploration in Reinforcement Learning (SIERL), a novel method that actively guides exploration by setting sub-goals based on the agent’s learning progress. At the beginning of each episode, SIERL chooses a sub-goal from the frontier (the boundary of the agent’s known state space), before the agent continues exploring toward the main task objective. The key contribution of our method is the sub-goal selection mechanism, which provides state-action pairs that are neither overly familiar nor completely novel, assuring that the frontier is expanded systematically and that the agent is capable of reaching any state within it.

Learning from demonstration has proved itself useful for teaching robots complex skills with high sample efficiency. However, teaching long-horizon tasks with multiple skills is challenging as deviations tend to accumulate, the distributional shift becomes more evident, and human teachers become fatigued over time, thereby increasing the likelihood of failure. To address these challenges, we introduce (ST)^2, a sequential method for learning long-horizon manipulation tasks that allows users to control teaching flow by specifying keypoints, enabling structured and incremental demonstrations. Using this framework, we study how users respond to two teaching paradigms: a traditional monolithic approach in which users demonstrate the entire task trajectory at once, and a sequential approach in which the task is segmented and demonstrated step by step.

We propose ASkDAgger, a framework for active skill-level data aggregation in imitation learning that balances human query effort and learning efficiency through adaptive gating and prioritized replay. Demonstrated on language-conditioned robotic tasks.

Reinforcement learning (RL) is a promising approach for robotic manipulation, but it can suffer from low sample efficiency and requires extensive exploration of large state-action spaces. Recent methods leverage the commonsense knowledge and reasoning abilities of large language models (LLMs) to guide exploration toward more meaningful states. However, LLMs can produce plans that are semantically plausible yet physically infeasible, yielding unreliable behavior. We introduce LLM-TALE, a framework that uses LLMs’ planning to directly steer RL exploration. LLM-TALE integrates planning at both the task level and the affordance level, improving learning efficiency by directing agents toward semantically meaningful actions. Unlike prior approaches that assume optimal LLM-generated plans or rewards, LLM-TALE corrects suboptimality online and explores multimodal affordance-level plans without human supervision.

We introduce ExploRLLM, a method that guides reinforcement learning exploration using large language model priors. It accelerates convergence and improves performance in robot manipulation tasks with limited data.

This paper presents a framework for fast and robust motion planning designed to facilitate automated driving. The framework allows for real-time computation even for horizons of several hundred meters and thus enabling automated driving in urban conditions. This is achieved through several features. Firstly, a convenient geometrical representation of both the search space and driving constraints enables the use of classical path planning approach. Thus, a wide variety of constraints can be tackled simultaneously (other vehicles, traffic lights, etc.). Secondly, an exact cost-to-go map, obtained by solving a relaxed problem, is then used by A*-based algorithm with model predictive flavour in order to compute the optimal motion trajectory. The algorithm takes into account both distance and time horizons. The approach is validated within a simulation study with realistic traffic scenarios. We demonstrate the capability of the algorithm to devise plans both in fast and slow driving conditions, even when full stop is required.

Closed-loop validation of autonomous vehicles is an open problem, significantly influencing development and adoption of this technology. The main contribution of this paper is a novel approach to reproducible, scenario-based validation that decouples the problem into several sub-problems, while avoiding to brake the crucial couplings. First, a realistic scenario is generated from the real urban traffic. Second, human participants, drive in a virtual scenario (in a driving simulator), based on the real traffic. Third, human and automated driving trajectories are reproduced and compared in the real vehicle on an empty track without traffic. Thus, benefits of automation with respect to safety, efficiency and comfort can be clearly benchmarked in a reproducible manner. Presented approach is used to benchmark performance of SBOMP planner in one scenario and validate SuperHuman driving performance.

To achieve optimal robot behavior in dynamic scenarios we need to consider complex dynamics in a predictive manner. In the vehicle dynamics community, it is well know that to achieve time-optimal driving on low friction surface, the vehicle should utilize drifting. Hence, many authors have devised rules to split circuits and employ drifting on some segments. These rules are suboptimal and do not generalize to arbitrary circuit shapes (e.g., S-like curves). So, the question “When to go into which mode and how to drive in it?” remains unanswered. To choose the suitable mode (discrete decision), the algorithm needs information about the feasibility of different modes (continuous motion). This makes it a class of Task and Motion Planning (TAMP) problems, which are known to be hard to solve optimally in real-time. In the AI planning community, search methods are commonly used. However, they cannot be directly applied to TAMP problems due to the continuous component. Here, we present a search-based method that effectively solves this problem and efficiently searches in a highly dimensional state space with nonlinear and unstable dynamics. The space of the possible trajectories is explored by sampling different combinations of motion primitives guided by the search. Our approach allows to use multiple locally approximated models to generate motion primitives (e.g., learned models of drifting) and effectively simplify the problem without losing accuracy. The algorithm performance is evaluated in simulated driving on a mixed-track with segments of different curvatures (right and left). Our code is available at https://git.io/JenvB.

Intelligent manufacturing is becoming increasingly important due to the growing demand for maximizing productivity and flexibility while minimizing waste and lead times. This work investigates automated secondary robotic food packaging solutions that transfer food products from the conveyor belt into containers. A major problem in these solutions is varying product supply which can cause drastic productivity drops. Conventional rule-based approaches, used to address this issue, are often inadequate, leading to violation of the industry’s requirements. Reinforcement learning, on the other hand, has the potential of solving this problem by learning responsive and predictive policy, based on experience. However, it is challenging to utilize it in highly complex control schemes. In this paper, we propose a reinforcement learning framework, designed to optimize the conveyor belt speed while minimizing interference with the rest of the control system. When tested on real-world data, the framework exceeds the performance requirements (99.8% packed products) and maintains quality (100% filled boxes). Compared to the existing solution, our proposed framework improves productivity, has smoother control, and reduces computation time.