Abstract Although legged locomotion over a moderately rugged terrain can be accomplished by employing simple reactions to the ground contact information, more effective approach, which allows predictively avoiding obstacles, requires model of environment and control algorithm that takes this into account when planning footsteps leg movements. This article addresses issues perception modeling foothold selection in walking robot. An integrated system is presented robot traverse previously...

Achieving full autonomy in a mobile robot requires combining robust environment perception with onboard sensors, efficient mapping, and real‐time motion planning. All these tasks become more challenging when we consider natural, outdoor that has many degrees of freedom (DOF). In this paper, address the issues planning legged walking over rough terrain, using only its sensors to gather necessary model. The proposed solution takes limited perceptual capabilities into account. A multisensor...

This paper considers motion planning for a six-legged walking robot in rough terrain, considering both the geometry of terrain and its semantic labeling. The labels allow to distinguish between different types surfaces it can walk on, identify areas that cannot be negotiated due their physical nature. proposed environment map provides planner information about shape class labels. Such as "wall" "plant" denote have avoided, whereas other labels, "grass", "sand", "concrete", etc. represent...

Scene and object reconstruction is an important problem in robotics, particular planning collision-free trajectories or manipulation. This paper compares two strategies for the of nonvisible parts surface from a single RGB-D camera view. The first method, named DeepSDF predicts Signed Distance Transform to given point 3D space. second MirrorNet reconstructs occluded objects' by generating images other side observed object. Experiments performed with objects ShapeNet dataset, show that...

A biologically inspired approach to feasible gait learning for a hexapod robot The objective of this paper is develop patterns that could be used control real walking robot. These gaits should enable the fastest movement possible with given robot's mechanics and drives on flat terrain. Biological inspirations are commonly in design robots their algorithms. However, legged differ significantly from biological counterparts. Hence we believe learned using or its simulation model rather than...

This paper concerns the problem of how to learn grasp dexterously, so as be able then novel objects seen only from a single viewpoint. Recently, progress has been made in data-efficient learning generative models that transfer well objects. These are learned demonstration (LfD). One weakness is that, this shall show, under challenging single-view conditions unreliable. Second, number model elements increases linearly training examples. This, turn, limits potential these for generalization...

This paper addresses the issues of unstructured terrain modelling for purpose motion planning in an insect-like walking robot. Starting from well-established elevation grid concept adopted to specific requirements a small legged robot with limited perceptual capabilities we extend this idea multi-sensor system consisting 2D pitched laser scanner, and stereo vision camera. We propose extension usual, ad hoc update mechanism by incorporating formal treatment spatial uncertainty, show how data...

Abstract This paper addresses the issues of unstructured terrain modeling for purpose navigation with legged robots. We present an improved elevation grid concept adopted to specific requirements a small robot limited perceptual capabilities. propose extension update mechanism by incorporating formal treatment spatial uncertainty. Moreover, this presents uncertainty models structured light RGB-D sensor and stereo vision camera used produce dense depth map. The model is based on propagation...

Many recent solutions to the RGB-D SLAM problem use pose-graph optimization approach, which marginalizes out actual depth measurements. In this paper we employ same type of factor graph optimization, but investigate gains coming from maintaining a map RGBD point features and modeling spatial uncertainty these features. We demonstrate that accuracy can be increased by employing models reflecting errors introduced measurements image processing. The new approach is validated in simulations...

This paper presents an adaptive foothold planning method for a hexapod walking robot. A local terrain map acquired with inexpensive structured light sensor is exploited as the information source algorithm, which uses polynomial-based approximation to create decision surface. The robot learns from simulations, therefore no priori knowledge required. results show that general enough work on various types of terrain. planned footholds enable walk more stable, avoiding slippages and fall-downs.

Purpose – The purpose of this paper is to describe a novel application the recently introduced concept from computer vision self‐localization walking robot in unstructured environments. technique described enables with monocular system (single camera) obtain precise estimates its pose regard six degrees freedom. This capability essential search and rescue missions collapsed buildings, polluted industrial plants, etc.

In this article, the design of a five-fingered anthropomorphic gripper is presented specifically designed for manipulation elastic objects. The manipulator features hybrid design, being equipped with three fully actuated fingers precise manipulation, and two underactuated, tendon-driven digits secure power grasping. For ease reproducibility, uses as many off-the-shelf 3D-printed components possible. on-board controller circuit firmware are also presented. includes resistive position angle...

The purpose of this paper is to propose and evaluate the method for grasping a defined set objects in an unstructured environment. To end, authors integrating convolutional neural network (CNN)-based object detection category-free method. considered scenario related mobile manipulating platforms that move freely between workstations manipulate objects. In application, robot not positioned with respect table manipulated detects environment uses methods determine reference pose gripper.,The...

This paper presents a posture optimization algorithm for six-legged walking robot. During on rough terrain and planning its motion the robot has to determine horizontal position, distance ground, inclination of platform. The proposed is based Particle Swarm Optimization method. increases stability margin maximizes possible range (by maximizing kinematic each leg). computation performed by using an analytical function obtained with Gaussian approximation. Gaussian-based approximation...

In this paper we present a traversability assessment method for motion planning in autonomous walking robots. The aim is to plan the of robot real scenario on rough terrain, where level details obtained terrain maps not sufficient planning. A guided RRT (Rapidly-exploring Random Trees) algorithm used terrain. We are looking that can learn cost function benefit guiding algorithm. probabilistic regression technique solve problem. Computing predictions values use planner explore space possible...

In this article, we propose a new method for constraints evaluation during the path planning of multi-legged walking robot. We application Gaussian Mixtures (GM) to build constraint models. With proposed analytical function can compute gradient and move robot out bound configuration. The allows checking self-collisions, workspace in few microseconds efficiently plan motion uses RRT-Connect framework. show how apply methods optimize posture robot, position feet swing phase explore search...