Localization algorithms that rely on 3D LiDAR scanners often encounter temporary failures due to various factors, such as sensor faults, dust particles, or drifting. These failures can result in a misalignment between the robot’s estimated pose and its actual position in the global map. To address this issue, the process of global re-localization becomes essential, as it involves accurately estimating the robot’s current pose within the given map. In this article, we propose a novel global re-localization framework that addresses the limitations of current algorithms heavily reliant on scan matching and direct point cloud feature extraction. Unlike most methods, our framework eliminates the need for an initial guess and provides multiple top-� candidates for selection, enhancing robustness and flexibility. Furthermore, we introduce an event-based re-localization trigger module, enabling autonomous robotic missions. Focusing on subterranean environments with low features, we leverage range image descriptors derived from 3D LiDAR scans to preserve depth information. Our approach enhances a state-of-the-art data-driven descriptor extraction framework for place recognition and orientation regression by incorporating a junction detection module that utilizes the descriptors for classification purposes. The effectiveness of the proposed approach was evaluated across three distinct real-life subterranean environments.

The article introduces an adaptive threshold detection mechanism aimed at enhancing a decentralized radar inertial odometry (RIO) framework, enabling resilient localization in challenging hostile environments. Considering that frequency-modulated continuous-wave (FMCW) radars possess characteristics enabling measurements in challenging indoor and outdoor environments, the fusion of multiple radars in an ensemble configuration, along with an inertial measurement unit (IMU), holds promise in surpassing individual sensor limitations. This approach thereby enhances robust perception. The proposed adaptive autonomous residual threshold mechanism employs a real-time residual analysis to dynamically adjust the sensor fusion process by comparing the variance between two extended Kalman filters. This adaptive approach addresses irregularities in data samples from multiple sensors, thereby enhancing the decentralized smoothing estimator’s precision in providing localization while navigating through hostile environments marked by limited visibility, extreme weather, or high interference. Consequently, it contributes to the resilience and adaptability of autonomous systems in real-world scenarios. The proposed framework effectively showcases precise localization through decentralized radar inertial odometry(RIO).

In this article, we propose a body-aware local navigation strategy for asymmetric holonomic robots for collision-free navigation in narrow pathways with sharp turns. In such scenarios, a robot with non-circular or asymmetric footprint that is comparable to the dimension of the pathways collides with walls when tracking Voronoi paths or risk-aware paths. This problem is addressed in this article through a novel multi-control barrier functions (CBF) based control strategy that achieves the objective of safe collision-free maneuvering at sharp turns. The proposed method is significantly computationally light in comparison to approaches based on model predictive control and online occupancy-grid based free-space and collision detection. In the proposed approach, a minimal set of parameters that characterize a sharp turn and the robot footprint are used to define six control barrier functions that define safe and unsafe regions of operation for a robot. A quadratic programming based CBF safety filter is designed that takes a nominal goal-reaching control as input and returns a minimally-deviating output that enforces the control barrier constraints and renders the safe set forward invariant throughout the turning maneuver. The three kinematic control inputs of the holonomic robot are shared in a conflict-free manner among the six control barrier constraints. The proposed local navigation approach was thoroughly validated in multiple scenarios in a simulated environment, where a robot with asymmetric footprint achieves collision-free maneuvering along multiple sharp turns, while respecting the safety and actuation constraints.

AMBITIOUS [1] aims at providing a fundamental technological infrastructure, which will offer advanced data aggregation and clean-up, analytics, AI-enabled forecasting and secure information exchange mechanisms, via a transparent computing continuum infrastructure, to be integrated with existing, mature services of SMEs, unleashing for them yet unforeseen functionalities and opening up new pathways of commercial exploitation. The envisaged fundamental infrastructure will be provided via the deployment of technological pillars, which will interact with existing services towards supporting the envisioned functionalities. The purpose of a pillar is to provide the same generic functionalities to diverse services, demonstrated by a various set of specific use cases, as a concrete processing chain, aiming at avoiding unnecessary redundancy of resources and budget. This paper presents the focused use cases.

Multiple modern robotic applications benefit from centralized cognition and processing schemes. However, modern equipped robotic platforms can output a large amount of data, which may exceed the capabilities of modern wireless communication systems if all data is transmitted without further consideration. This research presents a multi-agent, centralized, and real-time 3D point cloud map merging scheme for ceaselessly connected robotic agents. Centralized architectures enable mission awareness to all agents at all times, making tasks such as search and rescue more effective. The centralized component is placed on an edge server, ensuring low communication latency, while all agents access the server utilizing a fifth-generation (5G) network. In addition, the proposed solution introduces a communication-aware control function that regulates the transmissions of map instances to prevent the creation of significant data congestion and communication latencies as well as address conditions where the robotic agents traverse in limited to no coverage areas. The presented framework is agnostic of the used localization and mapping procedure, while it utilizes the full power of an edge server. Finally, the efficiency of the novel established framework is being experimentally validated based on multiple scenarios.

In this article, a novel approach for merging 3-D point cloud maps in the context of egocentric multirobot exploration is presented. Unlike traditional methods, the proposed approach leverages state-of-the-art place recognition and learned descriptors to efficiently detect overlap between maps, eliminating the need for the time-consuming global feature extraction and feature matching process. The estimated overlapping regions are used to calculate a homogeneous rigid transform, which serves as an initial condition for the general iterative closest point (GICP) point cloud registration algorithm to refine the alignment between the maps. The advantages of this approach include faster processing time, improved accuracy, and increased robustness in challenging environments. Furthermore, the effectiveness of the proposed framework is successfully demonstrated through multiple field missions of robot exploration in a variety of different underground environments.

In the field of resource-constrained robots and the need for effective place recognition in multi-robotic systems, this article introduces RecNet, a novel approach that concurrently addresses both challenges. The core of RecNet’s methodology involves a transformative process: it projects 3D point clouds into range images, compresses them using an encoder-decoder framework, and subsequently reconstructs the range image, restoring the original point cloud. Additionally, RecNet utilizes the latent vector extracted from this process for efficient place recognition tasks. This approach not only achieves comparable place recognition results but also maintains a compact representation, suitable for sharing among robots to reconstruct their collective maps. The evaluation of RecNet encompasses an array of metrics, including place recognition performance, the structural similarity of the reconstructed point clouds, and the bandwidth transmission advantages, derived from sharing only the latent vectors. Our proposed approach is assessed using both a publicly available dataset and field experiments 1 confirming its efficacy and potential for real-world applications.

This article establishes the novel D_+^*, , a risk-aware and platform-agnostic heterogeneous global path planner for robotic navigation in complex environments. The proposed planner addresses a fundamental bottleneck of occupancy-based path planners related to their dependency on accurate and dense maps. More specifically, their performance is highly affected by poorly reconstructed or sparse areas (e.g. holes in the walls or ceilings) leading to faulty generated paths outside the physical boundaries of the 3-dimensional space. As it will be presented, D+* addresses this challenge with three novel contributions, integrated into one solution, namely: (a) the proximity risk, (b) the modeling of the unknown space, and (c) the map updates. By adding a risk layer to spaces that are closer to the occupied ones, some holes are filled, and thus the problematic short-cutting through them to the final goal is prevented. The novel established D₊*  also provides safety marginals to the walls and other obstacles, a property that results in paths that do not cut the corners that could potentially disrupt the platform operation. D₊*  has also the capability to model the unknown space as risk-free areas that should keep the paths inside, e.g in a tunnel environment, and thus heavily reducing the risk of larger shortcuts through openings in the walls. D₊* is also introducing a dynamic map handling capability that continuously updates with the latest information acquired during the map building process, allowing the planner to use constant map growth and resolve cases of planning over outdated sparser map reconstructions. The proposed path planner is also capable to plan 2D and 3D paths by only changing the input map to a 2D or 3D map and it is independent of the dynamics of the robotic platform. The efficiency of the proposed scheme is experimentally evaluated in multiple real-life experiments where D₊* is producing successfully proper planned paths, either in 2D in the use case of the Boston dynamics Spot robot or 3D paths in the case of an unmanned areal vehicle in varying and challenging scenarios.