Online motion planning is a challenging problem for intelligent robots moving in dense environments with dynamic obstacles, e.g., crowds. In this work, we propose novel approach optimal and safe online minimal information about obstacles. Specifically, our requires only the current position of obstacles their maximum speed, but it does not need any exact trajectories or model. The proposed methodology combines Monte Carlo Tree Search (MCTS), via model simulations, Velocity Obstacles (VO),...

Partially Observable Markov Decision Processes (POMDPs) are a powerful framework for planning under uncertainty. They allow to model state uncertainty as belief probability distribution. Approximate solvers based on Monte Carlo sampling show great success relax the computational demand and perform online planning. However, scaling complex realistic domains with many actions long horizons is still major challenge, key point achieve good performance guiding action-selection process...

Dynamic Movement Primitives (DMPs) are a framework for learning trajectory from demonstration. The can be learned efficiently after only one demonstration, and it is immediate to adapt new goal positions time duration. Moreover, the also robust against perturbations. However, obstacle avoidance DMPs still an open problem. In this work, we propose extension of support volumetric based on use superquadric potentials. We show advantages approach when obstacles have known shape, extend unknown...

Autonomy is the frontier of research in robotic surgery and its aim to improve quality surgical procedures next future. One fundamental requirement for autonomy advanced perception capability through vision sensors. In this paper, we propose a novel calibration technique scenario with da Vinci robot. Calibration camera robot necessary precise positioning tools order emulate high performance surgeons. Our tailored RGB-D camera. Different tests performed on relevant use cases prove that...

Robotic surgery has significantly improved the quality of surgical procedures. In past, researches have been focused on automating simple actions. However, there exists no scalable framework for automation in surgery. this paper, we present a knowledge-based modular articulated tasks, example, with multiple coordinated The is consisted ontology, providing entities and rules task planning, "dynamic movement primitives" as adaptive motion planner to replicate dexterity surgeons. To validate...

The use of robots in minimally invasive surgery has improved the quality standard surgical procedures. So far, only automation simple actions been investigated by researchers, while execution structured tasks requiring reasoning on environment and choice among multiple is still managed human surgeons. In this paper, we propose a framework to implement task automation. consists task-level module based answer set programming, low-level motion planning dynamic movement primitives, situation...

Obstacle avoidance for DMPs is still a challenging problem. In our previous work, we proposed framework obstacle based on superquadric potential functions to represent volumes. this extend work include the velocity of trajectory in definition potential. Our formulations guarantee smoother behavior with respect state-of-the-art point-like methods. Moreover, new formulation allows obtain proximity than when using static (i.e. independent) We validate simulated multi-robot scenario and...

Over the last decade, use of robots in production and daily life has increased. With increasingly complex tasks interaction different environments including humans, are required a higher level autonomy for efficient deliberation. Task planning is key element It combines elementary operations into structured plan to satisfy prescribed goal, given specifications on robot environment. In this manuscript, we present survey recent advances application logic programming problem task planning....

Abstract Natural language annotations and manuals can provide useful procedural information relations for the highly specialized scenario of autonomous robotic task planning. In this paper, we propose publicly release AUTOMATE, a pipeline automatic knowledge extraction from expert-written domain texts. AUTOMATE integrates semantic sentence classification, role labeling, identification connectors, in order to extract templates Linear Temporal Logic (LTL) that be directly implemented any...

Robot-assisted surgery is an established clinical practice. The automatic identification of surgical actions needed for a range applications, including performance assessment trainees and process modeling autonomous execution monitoring. However, supervised action not feasible, due to the burden manually annotating recordings potentially complex long executions. Moreover, often few example executions procedure can be recorded. This paper proposes novel fast algorithm unsupervised in standard...

Abstract The quality of robot-assisted surgery can be improved and the use hospital resources optimized by enhancing autonomy reliability in robot’s operation. Logic programming is a good choice for task planning because it supports reliable reasoning with domain knowledge increases transparency decision making. However, prior typically incomplete, often needs to refined from executions surgical task(s) under consideration avoid sub-optimal performance. In this paper, we investigate...

Partially Observable Markov Decision Processes (POMDPs) are a powerful framework for planning under uncertainty. They allow to model state uncertainty as belief probability distribution. Approximate solvers based on Monte Carlo sampling show great success relax the computational demand and perform online planning. However, scaling complex realistic domains with many actions long horizons is still major challenge, key point achieve good performance guiding action-selection process...

Autonomous robotic surgery requires deliberation, i.e. the ability to plan and execute a task adapting uncer-tain dynamic environments. Uncertainty in surgical domain is mainly related partial pre-operative knowledge about patient-specific anatomical properties. In this paper, we introduce logic-based framework for tasks with deliberative functions of monitoring learning. The DE-liberative Framework Robot-Assisted Surgery (DEFRAS) estimates plan, executes it while continuously measuring...

Autonomy in robotic surgery will significantly improve the quality of interventions terms safety and recovery time for patient, reduce fatigue surgeons hospital costs. A key requirement such autonomy is ability surgical system to encode reason with commonsense task knowledge, adapt variations introduced by scenarios individual patients. However, it difficult all variability anatomy patients a priori, new knowledge often needs be acquired merged existing knowledge. At same time, not possible...

Autonomy in robot-assisted surgery is essential to reduce surgeons' cognitive load and eventually improve the overall surgical outcome. A key requirement for autonomy a safety-critical scenario as lies generation of interpretable plans that rely on expert knowledge. Moreover, Autonomous Robotic Surgical System (ARSS) must be able reason dynamic unpredictable anatomical environment, quickly adapt plan case unexpected situations. In this paper, we present modular Framework Robot-Assisted...

Partially Observable Monte Carlo Planning (POMCP) is an efficient solver for Markov Decision Processes (POMDPs). It allows scaling to large state spaces by computing approximation of the optimal policy locally and online, using a Tree Search based strategy. However, POMCP suffers from sparse reward function, namely, rewards achieved only when final goal reached, particularly in environments with long horizons. Recently, logic specifications have been integrated into guide exploration satisfy...

Hyper-redundant Robotic Manipulators (HRMs) offer great dexterity and flexibility of operation, but solving Inverse Kinematics (IK) is challenging. In this work, we introduce VO-FABRIK, an algorithm combining Forward Backward Reaching (FABRIK) for repeatable deterministic IK computation, approach inspired from velocity obstacles to perform path planning under collision joint limits constraints. We show preliminary results on industrial HRM with 19 actuated joints. Our achieves good...

In this paper, a dynamic model for an artificial finger driven by Shape Memory Alloy (SMA) wires is presented. Due to their high energy density, these alloys permit the realization of highly compact actuation solutions with potential applications in many areas robotics, ranging from industrial biomedical ones. Despite advantages, SMAs exhibit nonlinear and hysteretic behavior which complicates system design, modeling, control. case SMA are used activate complex robotic systems, further...

In recent years, the feasibility of autonomy in field Robotic-Assisted Minimally Invasive Surgery (RAMIS) has been investigated. One most important requirements for such a system is capability reconstructing patient's 3D anatomy real-time and registering it with pre-operative data. This crucial step surgical guidance augmented reality applications. A common solution to use Simultaneous Localization Mapping (SLAM) which plays an role computer vision robotics. Conventional SLAM algorithms...

Colonoscopy is the gold standard examination procedure for screening colorectal cancer (CRC), providing ability to inspect lower digestive tract [1]. Standard colonoscopy carried out with a minimally invasive approach based on flexible endoscope controlled by steering knobs. However, routine screening, intrusive nature of intervention, rigidity device and lack intuitiveness in can make experience highly unpleasant recipient [2]. Pa- tient discomfort pain are due tissue stretching friction as...