Skip to content

Commit

Permalink
Update index.md
Browse files Browse the repository at this point in the history
  • Loading branch information
Hi-Zed authored Feb 24, 2025
1 parent 14186c5 commit c635dc2
Showing 1 changed file with 4 additions and 21 deletions.
25 changes: 4 additions & 21 deletions index.md
Original file line number Diff line number Diff line change
Expand Up @@ -2,27 +2,10 @@
layout: default
---

Autonomous Robots are expected to carry out tasks that are inconvenient or unsafe for humans, such as inspecting critical infrastructure or exploring hostile environments. Prominent examples of such tasks include underwater and space exploration, nuclear decommissioning, underground tunnel inspection, and search and rescue operations. To deploy robots effectively in real-world environments, many challenges have to be addressed that concern the robots’ sensemaking, planning, decision-making and deliberation capabilities.
Autonomous robots are increasingly relied upon to perform tasks in outdoor environments that are challenging, inconvenient, or unsafe for humans. From exploring remote natural terrains to surveying disaster-stricken areas, these robots are expected to operate in settings that demand adaptability and robustness to change and uncertainty. Prominent examples include search and rescue operations, autonomous inspection and maintenance of outdoor plants, agricultural automation, and crop monitoring.
Effectively deploying robots in real-world outdoor environments presents numerous challenges related to their sensemaking, planning, decision-making, and deliberation capabilities. Outdoor operational scenarios are often characterized by unpredictable terrain, dynamic weather conditions, and limited access to reliable communication and sensing infrastructure. These environments introduce high levels of uncertainty, frequent changes, and increased risks of hardware failure, due to, for instance, exposure to extreme temperatures, moisture, or physical impacts from rugged terrain. In particularly extreme scenarios, robots are expected to complete their tasks before their hardware is compromised, e.g., due to radiation or other hostile environmental conditions.
In outdoor robotics, autonomy is particularly critical, as systems often need to function independently for extended periods without direct human intervention to successfully complete long-horizon tasks. This is especially true for scenarios where real-time human oversight is impractical, such as in the exploration of remote areas or even hostile environments characterized by adverse climates. However, many existing robotic systems remain either overly specialized for specific outdoor tasks or perform well only in constrained and simulated environments. While laboratory and simulation studies offer valuable insights, a significant gap persists between controlled experiments and the realities of deploying robots in unconstrained outdoor settings.

Real-world operational scenarios are often characterised by high levels of uncertainty and frequent changes. Limited sensor availability and increased risk of system failures are common issues in challenging operational settings. In particularly extreme scenarios, robots are expected to complete their tasks before their hardware is compromised, e.g., due to radiation or other hostile environmental conditions. The fact that field robots often need to operate autonomously for extended periods further exacerbates these challenges.

As a result, current systems are either overly specialised for particular tasks, or generalise only in constrained and simulated environments. While laboratory and simulation studies can provide valuable insights, the gap between experiments in artificial environments and unconstrained real-world settings remains significant. Consequently, representative datasets and benchmarking methods are necessary to bridge this gap.

The objective of this workshop is twofold. Firstly, it aims to encourage research that showcases the deployment of autonomous robots in unconstrained real-world environments. Secondly, it aims to promote the proposal of novel benchmarks for validating the actual capabilities of these systems in the field. This workshop will bring together experts from a range of disciplines within Robotics and Artificial Intelligence, to discuss the latest research and technological developments in autonomous field robotics.

**Topics of interest** include, but are not limited to:

**Navigation and mapping**, comprising algorithms for mapping and localization, path planning, and obstacle avoidance for navigation. Particular emphasis will be given to works tackling long-term autonomy and navigation in uncontrolled environments, under unpredictable and possibly changing working conditions.

**Perception and sensor fusion**, effective and robust sensory processing is one key prerequisite for the successful deployment of robots in unconstrained environments. The workshop welcomes contributions focused on the challenges of interpreting, integrating, and managing real-world data, especially in the context of operational settings where sensor availability is restricted - e.g., underground and underwater settings.

**Planning, Reasoning and decision-making**, another important requirement of robot deployment in real-world environments is the improvement of the robots’ capability to plan their actions and make decisions under varying conditions and under uncertainty. Different reasoning capabilities are required to this aim, including but not limited to physics, spatial and temporal reasoning. In the context of critical applications, such as search and rescue and infrastructure maintenance tasks, it is also crucial for robots to be capable of explaining their reasoning steps and subsequent actions.

**Improved Knowledge Representations**. Designing representations that can adequately express the heterogeneity of the real world, that can be modularly adapted to model environmental changes, and that reconcile the various knowledge resources that contribute to the reasoning process (e.g., domain-specific, task-specific, commonsense-driven, factual/encyclopaedic knowledge) is an open challenge in Robotics.

**Real-world datasets**, which capture the high variability of real-world environments, as well as the unpredictability of environmental conditions. The availability of such heterogeneous datasets contributes to more rigorous and truthful benchmarking practices.

Overall, the workshop aims to provide a forum for researchers and practitioners to exchange ideas, discuss challenges and opportunities, and identify new research directions to bridge the gap between laboratory experiments and real-world deployment.
By bringing together experts from a range of disciplines, this workshop will facilitate cross-disciplinary knowledge transfer and exchange towards the development of innovative robotic applications that can produce a real-world impact.
This workshop welcomes contributions aimed at addressing this gap, to foster a multi-disciplinary discussion of the challenges, implications, and objectives of deploying autonomous robots to support outdoor tasks in demanding scenarios.


0 comments on commit c635dc2

Please sign in to comment.