We present an efficient asymptotically-optimal randomized motion planning algorithm solving single-query path planning problems using a bidirectional search. The algorithm combines the benefits from the widely known algorithms RRT-Connect and RRT∗ and scores better than both by finding a solution faster than RRT∗, and -unlike RRT-Connect — converging towards a theoretical optimum. We outline the proposed algorithm and proof its optimality. The efficiency and robustness is demonstrated in a number of real world applications which benefit from the bidirectional approach: planning car trajectories in a parking garage for the autonomous vehicle CoCar, generating cost-efficient trajectories for the multi-legged walking robot LAURON V in a planetary exploration scenario and performing mobile manipulation tasks for our highly actuated service robot HoLLiE. Moreover, we compare and show the improvements over "vanilla" RRT in a set of challenging benchmarks. RRT∗-Connect will contribute to increase the performance of autonomous robots and vehicles due to the reduced motion planning time in complex environments.

RRT∗-Connect: Faster, asymptotically optimal motion planning

In the near future, humans will be relieved from parking. Major improvements in autonomous driving allow the realization of automated valet parking (AVP). It enables the vehicle to drive to a parking spot and park itself. This paper presents a review of the intelligent vehicles literature on AVP. An overview and analysis of the core components of AVP such as the platforms, sensor setups, maps, localization, perception, environment model, and motion planning is provided. Leveraging the potential of AVP, high density parking (HDP) is reviewed as a future research direction with the capability to either reduce the necessary space for parking by up to 50 % or increase the capacity of future parking facilities. Finally, a synthesized view discussing the remaining challenges in automated valet parking and the technological requirements for high density parking is given.

The future of parking: A survey on automated valet parking with an outlook on high density parking

This work provides an architecture that incorporates depth and tactile information to create rich and accurate 3D models useful for robotic manipulation tasks. This is accomplished through the use of a 3D convolutional neural network (CNN). Offline, the network is provided with both depth and tactile information and trained to predict the object’s geometry, thus filling in regions of occlusion. At runtime, the network is provided a partial view of an object. Tactile information is acquired to augment the captured depth information. The network can then reason about the object’s geometry by utilizing both the collected tactile and depth information. We demonstrate that even small amounts of additional tactile information can be incredibly helpful in reasoning about object geometry. This is particularly true when information from depth alone fails to produce an accurate geometric prediction. Our method is benchmarked against and outperforms other visual-tactile approaches to general geometric reasoning. We also provide experimental results comparing grasping success with our method.

Multi-Modal Geometric Learning for Grasping and Manipulation

—This paper gives an overview on our frameworkfor efﬁcient collision detection in robotic applications. It uniﬁesdifferent data structures and algorithms that are optimizedfor Graphics Processing Unit (GPU) architectures. A speed-upin various planning scenarios is achieved by utilizing storagestructures that meet speciﬁc demands of typical use-cases likemobile platform planning or full body planning. The system isalso able to monitor the execution of motion trajectories forintruding dynamic obstacles and triggers a replanning or stopsthe execution. The presented collision detection is deployed inlocal dynamic planning with live pointcloud data as well as inglobal a-priori planning. Three different mobile manipulationscenarios are used to evaluate the performance of our approach. I. I NTRODUCTION Complex autonomous systems typically rely on time in-tensive and non-agile motion planning software to achievegiven objectives. This depicts a big handicap which needsto be overcome to bring mobile service robots into reallife scenarios. We tackle the problem through massivelyparallelizing the collision detection sub-problem to solve iton General Purpose GPU (GPGPU) Hardware.In this work we ﬁrst introduce the characteristics ofplanning data and describe the capabilities of our uniﬁedcollision detection framework before we illustrate its practi-cal application by the example of motion primitive and fullbody planning for our mobile manipulation platforms fromFig. 7.II. R

Unified GPU voxel collision detection for mobile manipulation planning.

Skill transfer learning for autonomous robots and human-robot cooperation: A survey

In order to perform useful tasks, a service robot needs to manipulate objects in its environment. In this paper, we propose a method for experimental evaluation of the suitability of a robotic hand for grasping tasks in service robotics. The method is applied to the Schunk 5-Finger Gripping Hand, which is a mechatronic gripper designed for service robots. During evaluation, it is shown, that it is able to grasp various common household objects and execute the grasps from the well known Cutkosky grasp taxonomy [1]. The result is, that it is a suitable hand for service robot tasks.

/pdf/experimental-evaluation-of-the-schunk-5-finger-gripping-hand-3mlce9ezhy.pdf

Experimental evaluation of the schunk 5-Finger gripping hand for grasping tasks

Proactive collision detection enables robots to efficiently execute tasks in shared human-robot-workspaces by avoiding collision-prone situations. Our work connects motion prediction of RGB-D flow algorithms with motion primitive planning via an efficient voxel Swept-Volume-based collision detection. The approach can handle scenarios with varying contents as we use the same techniques to predict single moving objects but also articulated bodies. Our process chain consists of highly parallel GPU algorithms that allow a full 3D representation of planned trajectories and predictions from live pointclouds in high resolution, while still being online capable. We demonstrated our achievements in two scenarios with different motion granularity.

Anticipate your surroundings: Predictive collision detection between dynamic obstacles and planned robot trajectories on the GPU

We present our recent work on the soft- and hardware design of the bimanual mobile manipulation platform HoLLiE that is equipped with an actuated upper body. The goal was to develop a robust but extensible robot with a non-intimidating abstract anthropomatic appearance based on a combination of industrial robotic components and intelligent mechatronics. With a range of different sensors and a highly articulated body HoLLiE can handle everyday objects, interact with humans in multiple ways and therefore be employed in various service robotic scenarios. We demonstrate the usability of our concept by quantifying the workspace and its stability and also briefly describe the software components.

Andreas Hermann

Papers

RRT∗-Connect: Faster, asymptotically optimal motion planning

Unified GPU voxel collision detection for mobile manipulation planning.

Experimental evaluation of the schunk 5-Finger gripping hand for grasping tasks

Anticipate your surroundings: Predictive collision detection between dynamic obstacles and planned robot trajectories on the GPU

Hardware and software architecture of the bimanual mobile manipulation robot HoLLiE and its actuated upper body