Recent advances in mobile robotics and AI promise to revolutionize industrial production. As autonomous robots are able to solve more complex tasks, the difficulty of integrating various robot skills and coordinating groups of robots increases dramatically. Domain independent planning promises a possible solution. For single robot systems a number of successful demonstrations can be found in scientific literature. However our experiences at the RoboCup Logistics League in 2017 highlighted a severe lack in plan quality when coordinating multiple robots. In this work we demonstrate how out of the box temporal planning systems can be employed to increase plan quality for temporal multi-robot tasks. An abstract plan is generated first and sub-tasks in the plan are auctioned off to robots, which in turn employ planning to solve these tasks and compute bids. We evaluate our approach on two planning domains and find significant improvements in solution coverage and plan quality.

In current approaches to combined task and motion planning, usually symbolic planning and sampling based motion-planning are integrated. One problem is here to come up with good samples. We address the problem of identifying useful poses for a robot close to working surfaces such as tables or shelves. Our approach is based on reachability inversion which answers the question: where should the robot be located in order to reach a certain object? We extend the concept from point-based objects to flat polygonal surfaces in order to enable the robot to have a a good grasping position for many objects. Our approach allows to quickly sample multiple distinct poses for the robot from an prior computed distribution. Further we show how sampling from an inverse reachability distribution can be integrated into a CTAMP system.

Searching for objects and observing parts of a known environment efficiently is a fundamental problem in many real‐world robotic applications, e.g., household robots searching for objects, inspection robots searching for leaking pipelines, and rescue robots searching for survivors after a disaster. We consider the problem of identifying and planning sequences of sensor locations from which robot sensors can observe and cover complex three‐dimensional (3D) environments while traveling only short distances. Our approach is based on sampling and ranking a large number of sensor locations for a 3D environment represented by an OctoMap. The visible area from these sensor locations induces a minimal partition of the 3D environment that we exploit for planning sequences of sensor locations with short travel times efficiently. We present multiple planning algorithms designed for single robots and for multirobot teams. These algorithms include variants that are greedy, optimal, or based on decomposing the planning problem into a set cover and traveling salesman problem. We evaluated and compared these algorithms empirically in simulation and real‐world robot experiments with up to four robots. Our results demonstrate that, despite the intractability of the overall problem, computing and executing effective solutions for multirobot coverage search in real 3D environments is feasible and ready for real‐world applications.

Domain-independent planning in general is broadly applicable to a wide range of tasks. Many formalisms exist that allow the description of different aspects of realistic problems. Which one to use is often no obvious choice, since a higher degree of expressiveness usually comes with an increased planning time and/or a decreased policy quality. Under the assumption that hard guarantees are not required, users are faced with a decision between multiple approaches. As a generic model we use a probabilistic description in the form of Markov Decision Processes (MDPs). We define abstracting translations into a classical planning formalism and fully observable nondeterministic planning. Our goal is to give insight into how state-of-the-art systems perform on different MDP planning domains.

To autonomously carry out complex mobile manipulation tasks, a robot control system has to integrate several components for perception, world modeling, action planning and replanning, navigation, and manipulation. In this paper, we present a modular framework that is based on the Temporal Fast Downward Planner and supports external modules to control the robot. This allows to tightly integrate individual sub-systems with the high-level symbolic planner and enables a humanoid robot to solve challenging mobile manipulation tasks. In the work presented here, we address mobile manipulation with humanoids in cluttered environments, particularly the task of collecting objects and delivering them to designated places in a home-like environment while clearing obstacles out of the way. We implemented our system for a Nao humanoid tidying up a room, i.e., the robot has to collect items scattered on the floor, move obstacles out of its way, and deliver the objects to designated target locations. Despite the limited sensing and motion capabilities of the low-cost platform, the experiments show that our approach results in reliable task execution by applying monitoring actions to verify object and robot states.

We present the integration of our symbolic planner as the high-level executive in the Tidyup-Robot project. Tidyup-Robot deals with mobile manipulation scenarios in a household setting. We introduce our system architecture and report on issues and advantages observed during development and deployment.

Task coordination for autonomous mobile service robots typically involves a substantial amount of background knowledge and explicit action sequences to acquire the relevant information nowadays. We strive for a system which, given a task, is capable of reasoning about task-relevant knowledge to automatically determine whether that knowledge is sufficient. If missing or uncertain, the robot shall decide autonomously on the actions to gain or improve that knowledge. In this paper we present our baseline system implementing the foundations for these capabilities. The robot has to analyze a tabletop scene and increase its object type confidence. It plans motions to observe the scene from multiple perspectives, combines the acquired data, and performs a recognition step on the merged input.

Although planning for the tasks a household robot has to perform appears to be easy, there exists the problem that the robot is usually uncertain about the state of the household when starting to plan. For example, when getting the order of tidying up the kitchen, the robot does not know what objects it will have to put away and whether there are actually any objects that need to be put away. Furthermore, while sensing operations can provide moreinformation about the environment, things can go wrong when executingan action. In this paper, we try to identify conditions under which classical planning can be used in a replanning loop in order to solve the planning problem in nondeterministic partially observable open domains. In particular, we will define completeness and soundness of replanning with respect to nondeterministic planning and we will identify a PSPACE-checkable condition that guarantees soundness.

State of the art classical planning systems can efficiently solve large symbolic problem instances. Applying classical planning techniques to robotics is possible by integrating geometric reasoning in the planning process. The problems that are solvable in this way are significantly smaller than purely logical formulations as many costly geometric calculations are requested by a planner. Therefore we aim to avoid those calculations while preserving correctness. We address this problem with efficient caching techniques. Subsumption caching avoids costly computations by caching geometric queries and beyond answering the same queries also considers less or more constrained ones. Additionally, we describe a lazy evaluation technique that pushes applicability checks for successor states performing geometric queries to a later point. As we are interested in the performance of our planner not as a standalone component, but as part of an intelligent robotic system, we evaluate those techniques embedded in an integrated system during real-world mobile manipulation experiments.

We present a search-based path planning system for ground robots on three dimensional terrain. Effectively negotiating such terrain often requires to utilize dedicated robot hardware and to execute specific behaviors. Our base system is independent from the actual robot configuration, but can be customized to a robot's abilities. We explicitly plan using a full 3d representation, not requiring any projection or slicing to a 2d world. The drivable surface manifold is automatically extracted from the volumetric 3d representation and generic motions are planned on these surface cells. This is achieved with behavior modules that integrate robot skills with the search. Such a behavior module is responsible for defining traversable surfaces, computing if a motion can be executed, and its cost. We implement two such modules: One for sloped ground and ramps, and one for steps and stairs. The approach is evaluated on simulated real-world environments.

The main goal of the paper is to continuously enlarge the set of software building blocks that can be reused in the search and rescue domain.

In recent years, domain-independent planning has been applied to a rising number of real-world applications. Usually, the description language of choice is PDDL. However, PDDL is not suited to model all challenges imposed by real-world applications. Dornhege et al. proposed semantic attachments to allow the computation of Boolean fluents by external processes called modules during planning. To acquire state information from the planning system a module developer must perform manual requests through a callback interface which is both inefficient and error-prone.In this paper, we present the Object-oriented Planning Language OPL, which incorporates the structure and advantages of modern object-oriented programming languages. We demonstrate how a domain-specific module interface that allows to directly access the planner state using object member functions is automatically generated from an OPL planning task. The generated domain-specific interface allows for a safe and less error-prone implementation of modules. We show experimentally that this interface is more efficient than the PDDL-based module interface of TFD/M.

This paper describes the software and hardware system developed by the University of Freiburg team of search and rescue robots for the Ro boCup Rescue 2010 competition. This system is an extension to the software that finishe d in first place the 2005 and 2006 autonomy challenge, focusing on two key areas: autonomous navigation and manipulation. Our team, consisting mainly of students, originates from the former CS Freiburg team (RoboCupSoccer) , th ResQ Freiburg team (RoboCupRescue Simulation), and RescueRobots Freibu rg teams ’05 and ’06.

This paper describes the approach of the RescueRobots Freiburg team, which is a team of students from the University of Freiburg that originates from the former CS Freiburg team (RoboCupSoccer) and ...

Autonomous vehicles with a car-like motion model, i.e. an Ackerman drive, can only follow trajectories with a smooth curvature. These trajectories are difficult to create with conventional planning algorithms. In this paper we present a method to create a spline-based trajectory from a given sequence of waypoints. We guess an initial trajectory and optimize it to minimize the time of travel. The resulting trajectory satisfies a number of constraints based on physical properties of the vehicle and the environment. Our experiments indicate that few iterations of our optimization algorithm are sufficient to produce valid and fast trajectories.