About the project

The AERO-TRAIN research and training programme goal is to enhance competitiveness of Europe in the sector of drones and I&M. The overall AERO-TRAIN programme has been conceived to create maximum value within the whole value chain of the sector, starting by boosting the career of new researchers, followed by increasing the knowledge and competence of partners and industry while keeping an eye on the society as very end users.

AERO-TRAIN exists to provide efficient and innovative inspection and maintenance tools for society and companies by providing a scientific background and training to the early-stage researchers in the field of Aerial Robotic Technologies for Inspection and Maintenance.

Goals:

• Researchers: To provide a scientific background to the early-stage researchers in the field of Aerial Robotic Technologies for Inspection and Maintenance.
• Society: Fill the gap between what can be done and has already been done (and inform the public of the findings and results.)
• Companies: Provide efficient inspection and maintenance tools.

Key Findings & Results

This section presents the primary research findings and results from the PhD students’ work. These findings represent advancements in both aerial robotics technology and human-robot interaction systems, focusing on novel methods of improving system performance, safety, and human-machine collaboration.

Aerial Robotics Innovations

A substantial portion (70%) of the PhD research contributed to enhancing the capabilities of aerial robotics related topics. These contributions include innovative designs, control systems, and methodologies aimed at improving the autonomy, safety, and efficiency of UAVs (Unmanned Aerial Vehicles) during inspection operations considering contact phases. Below are the key areas of innovation:

Novel Aerial Robot Designs

Some students optimized underactuated aerial vehicle designs by adjusting the center of mass to enhance manipulation, expanding UAV applications like non-destructive testing and dynamic scanning.

Others developed and tested platforms with tilting propellers for force exertion in complex directions.

These innovations transform UAVs from passive surveillance tools into interactive systems capable of complex physical tasks.

Advanced Control Systems for Aerial Manipulation

Several students developed hybrid force/position control systems enabling UAVs to handle heavy objects more effectively. These systems enhance stability during physical tasks, managing disturbances from surface contact and improving control of omnidirectional aerial robots.

Edge Computing for Autonomous Aerial Systems

Some research integrated edge computing with UAVs to offload heavy tasks like target detection and navigation, improving autonomy and performance by overcoming resource limits and network delays.

Human-Drone Interaction & Immersive Technologies

In parallel with innovations in aerial robotics, other PhD students focused on improving the interaction between humans and drones through enhanced technologies. Their work concentrated on simplifying the control of UAVs and improving the operator’s situational awareness.

Augmented Reality and Shared Decision-Making Systems

Students developed frameworks combining AR, deep learning, and shared decision-making to provide real-time visual overlays and assist operators in complex tasks, reducing stress and improving inspection performance.

Research on multimodal teleoperation integrated 3D visual and haptic feedback for better UAV control. Neural estimators and miniaturized haptic joysticks enabled precise quadrotor operation with minimal sensing.

Reinforcement Learning-Enhanced Algorithms

Reinforcement learning improved robot navigation, enhancing RRT* algorithms for efficient movement in complex environments, enabling robots to navigate autonomously while allowing human intervention when needed.

Light Field Displays for Immersive Interaction

Research on light field displays addresses 3D display limitations by enhancing depth perception and focus, improving situational awareness in remote aerial operations.

Advancements in human-drone interaction enable UAVs to handle complex tasks with human oversight, enhancing control, safety, and reducing cognitive load.

Neuromorphic Approaches for Civil Infrastructure Defect Detection

Neuromorphic, bio-inspired, and energy-efficient methods were used to detect civil infrastructure defects. Datasets from neuromorphic sensors, effective in low or dynamic lighting, were benchmarked with spiking neural networks. Detection accuracy, latency, and energy efficiency were evaluated on neuromorphic chips like Intel Loihi 2.

Impact & Benefits to the Public

The research carried out by the PhD students contributes directly to improving safety, efficiency, and reliability in infrastructure inspection and maintenance operations. These innovations have the potential to significantly enhance the way industries approach complex inspection tasks, reducing the need for human intervention in hazardous environments and increasing operational efficiency.

Enhancing Safety and Efficiency in Infrastructure Inspection

Students' research improves safety and reduces human involvement in hazardous infrastructure inspections by equipping UAVs with advanced control systems and interaction capabilities. These drones can perform tasks previously requiring human workers, minimizing risks and enhancing efficiency.

Several students developed UAVs for complex inspection and maintenance in industries like energy, transportation, and construction. These systems enable drones to interact physically with their environment, reducing worker injury risk while improving inspection speed and precision.

To improve operator experience, students developed haptic interfaces and control systems that provide real-time tactile feedback, such as a haptic finger joystick that simulates drone-environment interactions. This enhances situational awareness, reduces errors, and boosts inspection efficiency.

Shared decision-making algorithms and AI control systems reduce cognitive load on operators by automating routine tasks and assisting with precision, allowing more focus on high-level decisions and reducing fatigue.

Integrating augmented reality provides real-time visual overlays, improving operator understanding of the drone’s position and surroundings, leading to better decision-making and inspection outcomes.

In conclusion, the research carried out by the PhD students has direct implications for enhancing safety and operational efficiency in infrastructure inspection tasks. By reducing the need for human intervention in hazardous environments and improving the interaction between humans and UAVs, these innovations have the potential to transform how industries handle inspection and maintenance, resulting in safer, faster, and more cost-effective operations.

Future Outlook

As the PhD students get near the completion of their projects, they have provided insights into the future of their research and the potential impact it could have on industries such as infrastructure inspection. This section explores both the remaining challenges in aerial robotics and the students' personal plans for continuing their work, either in industry or academia.

Industrial Applications of Aerial Robotics

Many of the students intend to apply their research to industrial settings, focusing on the development and deployment of aerial robotics systems for infrastructure inspection. They foresee continued advancements in this field but acknowledge that there are still several challenges that need to be addressed before widespread adoption can occur.

Improving Robustness and Reliability

Future work aims to improve UAV reliability in dynamic environments, tackling challenges like high winds, variable payloads, and obstacles. Enhancing robustness is crucial for tasks like contact inspections and non-destructive testing (NDT) with minimal human intervention.

Adapting to Real-World Testing

While much research was done in simulations, students stressed the need for real-world testing. They aim to refine systems for infrastructure inspections by improving control algorithms to handle vibrations, obstacles, and interference.

Bringing New Technologies to Market

The transition from research to industry is a key priority for some students. They aim to commercialize their innovations by working with companies to integrate advanced UAV systems into existing infrastructure inspection frameworks. The goal is to make UAVs a standard tool in sectors such as energy, transportation, and construction, where they can reduce the need for manual inspections and enhance operational safety and efficiency.

Key Focus Areas for Industry:

• Developing adaptive control algorithms to improve UAV performance in unpredictable conditions.
• Enhancing safety protocols for operating UAVs in close proximity to infrastructure.
• Ensuring scalability and ease of integration into existing industrial workflows.

Continuous Development in Academia

A portion of the students plan to remain in academia or maintain close ties to academic research. They aim to build on their existing work, with a particular focus on advancing the technical capabilities of UAVs and exploring new applications of aerial robotics.

Expanding Research in Collaborative Aerial Robotics

Some students envision aerial robotics as collaborative systems, with multiple UAVs coordinating for tasks like lifting, positioning, and inspecting infrastructure. Future research will develop algorithms for seamless, autonomous multi-UAV cooperation.

Exploring Human-Robot Interaction and Immersive Technologies

Students aim to improve human-robot interaction in UAV operations for infrastructure inspections, using systems like haptic feedback, augmented reality, and shared autonomy. Their goal is to create more intuitive, immersive interfaces for safer, easier UAV management in industrial applications.

Academic-Industry Collaboration

Students value strong academia-industry connections and seek to continue collaborative research that bridges theoretical advancements with practical applications. They aim to scale their work by partnering with industry to address complex real-world challenges in aerial robotics.

Key Focus Areas for Academia:

• Researching collaborative UAV systems for large-scale infrastructure tasks.
• Enhancing operator interfaces to reduce cognitive load and improve precision in aerial manipulation.
• Continuing to explore multi-modal teleoperation frameworks for remote inspection tasks.

Future Challenges and Areas for Exploration

Both industry- and academia-focused students have identified several ongoing challenges that will need to be addressed in future research and development:

System Scalability and Integration

One challenge is ensuring that UAV systems can be easily integrated into existing infrastructure inspection processes. This involves creating systems that are not only highly effective but also cost-efficient, scalable, and compatible with current industrial workflows.

Real-World Performance Validation

As students continue to refine their research, they will focus on further validating system performance in diverse real-world environments. They emphasize the importance of safety testing and performance guarantees in challenging conditions, such as in high-altitude or high-vibration environments.

Addressing Regulatory and Operational Barriers

In addition to technical challenges, students foresee potential hurdles related to regulatory approval and operational constraints. Ensuring that UAV systems meet safety standards and compliance regulations for infrastructure inspection in various industries will be crucial to widespread adoption.

The future of the research conducted by these PhD students holds significant promise, both in terms of industrial application and further academic exploration.

In industry, students will focus on enhancing the robustness and reliability of UAV systems for infrastructure inspection, with a view to commercializing their technologies.

Those pursuing academic careers aim to continue developing innovative aerial robotics systems and exploring human-robot interaction.

Whether in industry or academia, the next phase of their work will involve addressing the remaining challenges in scalability, real-world validation, and regulatory compliance to ensure that these systems can be widely adopted and effectively implemented.