To Master Students: For interest in the proposed projects, contact us at: [email protected] 

Projects for Master Thesis - Open for academic period 2024-2025 

Rapid Wildfire Detection with a Team of Aerial Robots
Supervisors: Kostas Alexis, Tor Arne Johansen | Keywords: XPRIZE Wildfire Challenge
Available. 
This project aims to develop a sensing payload that enables the detection and localization of a wildfire virtually as soon as it is ignited, meaning at a very early stage of its evolution. This in turn shall allow rapid response to such potentially catastrophic events. The focus is both on the hardware development of the sensing load - which should be integrated on a small fixed-wing unmanned aerial vehicle - as well as on the robot perception and fusion algorithms for detection and localization from heights.  ​
Extinguishing Wildfires with Guided Airdrops
Supervisors: Kostas Alexis, Tor Arne Johansen | Keywords: XPRIZE Wildfire Challenge
Available. ​
This project aims to develop a specialized payload to be ferried by a fixed-wing unmanned aerial vehicle (UAV) that can be used to suppress a wildfire detected at its very early stages after ignition. The envisioned payload shall be released by the ferrying UAV at high speeds and using control surfaces (w/o propulsion) shall guide itself accurately to drop on the desired point of the fire area. Upon impact, the payload shall release a specialized fire retardant or other technology to eliminate the early-detected small wildfire area using a limited number of such airdrops. ​​
 
Extinguishing Wildfires with Launchable Airdrops
Supervisors: Kostas Alexis, Tor Arne Johansen | Keywords: XPRIZE Wildfire Challenge
Available. ​
This project aims to develop a specialized speed-launched unmanned aerial vehicle that integrates a fire retardant liquid (or other technology) such that it flies to crash on the area of an early-detected wildfire and suppresses it. The envisioned UAV shall accelerate based on its launching mechanism and use a battery and motor to further propel it before dropping this battery and then relying solely on a secondary smaller battery solely to actuate its control surfaces such that it can fly to the desired destination and "crash" the fire to suppress it. ​
  • Details available following this link.  
  • Project Relevance: XPRIZE Wildfire Challenge
  • Collaborating institutions: Brigham Young University
  • Main Contact: Kostas Alexis ([email protected])
Photo from source. 
 
Small and Fast VTOL Aerial Robot for Cluttered Flight
Supervisors: Kostas Alexis, Mihir Kulkarni | Keywords: Aerial Robotics
Available. ​
This project aims to develop a small-scale fast and agile quad-plane. The robot shall have a wingspan in the region of 1m and be able to fly fast and safely in an obstacle-filled environment such as a forest under the canopy. 
Photo from source. 
 
Enabling Safe Guidance in Perceptually Degraded Environments with mmWave Radars
Supervisors: Kostas Alexis, Morten Nissov | Keywords: Perception, SLAM
Available. ​
A methodology for safely following waypoints or acting upon user-commands is crucial for autonomy, and especially interesting in environments featuring challenging conditions. For example, autonomous agents relying on vision or LiDAR sensors will find great challenge in navigating through dense fog or smoke. Recently mmWave radar sensors have become a hot topic in robotics research for their ability to operate in conditions which prove difficult for typical perception sensors like vision and LiDAR. Research has demonstrated capable estimation methods, however there is little progress made in closing the loop. In these environments application of mmWave radar technology for guidance and obstacle avoidance can prove interesting as not only is the sensor unaffected by these degradations, but it also manages to directly measure environment-relative velocity, providing unique opportunities for control and obstacle avoidance which is less vulnerable to estimation breaking due to poor associations. The main goal of this project is to develop and implement a safe guidance law which takes advantage of the features of mmWave radar sensing technology, namely direct velocity measuring and robustness to certain environmental conditions.
Autonomous Docking for an Aerial and Legged Marsupial System
Supervisors: Kostas Alexis, Grzegorz Malczyk | Keywords: Learning for Robotics
Available. ​
Aerial robots are unconstrained by terrain, but they present limited endurance. Heterogeneous robot teams offer significant advantages for exploration and navigation missions in unknown environments. Notably, the air–ground robot teams have superior efficiency due to their mixed abilities and shared sensor resources, which is extremely useful when completing missions. However, the high energy consumption and reduced payload capability of the aerial platforms limit the complete robot team’s behaviour. At some point in a mission, the aircraft must return to the base, land, and replenish its batteries. This task restricts the team from exploring the environment in-depth and working for even longer durations. Therefore, this project aims to solve this problem and establish an autonomous system for docking. The on-site autonomous docking should happen multiple times, allowing the team to explore large areas and perform long-term missions by deploying the individual robots according to the current task-specific requirements.​
 
Learning and Evolution for Embodied Autonomy
Supervisors: Kostas Alexis | Keywords: Embodied Autonomy
Available
In September 2021, the finals of the DARPA Subterranean Challenge took place and Team CERBERUS – led by the PI – won the competition. Both for CERBERUS and for the other top-scoring teams, flying robots presented scouting capabilities but otherwise the bulk of the mission was carried by ground robots owing to their extended endurance and effective range. This contradicts our intuition regarding the benefits of flying robots in demanding subterranean environments with first being the fact that they do not depend on the challenges of the otherwise perilous terrain. A host of species, such as bats or insects, demonstrate the analogy of benefits of flying systems in natural underground environments but flying robots still need fundamental improvements to reach similar performance. This master thesis is motivated by the observation that despite the vastly different missions that flying robots undertake, and the different environments in which they operate, certain types of vehicles – primarily multirotors and fixed-wing systems – dominate across use cases. We claim that this is not only counterintuitive, as a higher variety of embodiments and autonomy solutions would follow the natural paradigm, but also we claim that this likely represents the key reason for the limitations observed on autonomous flying robots. Specifically, we claim that an environment-specific co-synthesis and co-optimization of the robot’s “body” (airframe, sensing, propulsion) and “brain” (autonomy functionality especially relying on data-driven learning methods for navigation) is necessary. Reflecting upon this analysis, in this master project you are tasked with employing the combined benefits of evolutionary algorithms and learning to facilitate the automated environment-specific adaptation of flying robot embodiments. 
Deep Reinforcement Learning for Embedded Control Policies for Aerial Vehicles
Supervisors: Kostas Alexis (NTNU), Mihir Kulkarni, Welf Rehberg Martin Jacquet | Keywords: Learning for Robotics
Available. ​
Unmanned aerial vehicles (UAVs) play an increasingly important role in industry and academia, being well suited for a wide variety of different tasks such as inspection and search and rescue missions. The area of small UAVs is dominated by designs developed from heuristics and empirical trial and error. While we employ sophisticated algorithms to generate optimal behavior strategies for these designs and don’t rely on human intuition, we fail to extend this to the design process. Employing algorithmic frameworks to jointly optimize over behavior and design can mitigate the suboptimality resulting from the human bias. Deep Reinforcement Learning is a promising method for generating near optimal behavior for arbitrary robot morphologies. However, its efficacy relies heavily on the physical fidelity of the training environment [1]. Since the behavior of aerial robots is influenced by a variety of dynamics that are challenging to accurately model, The proposed master project aims to enhance the precision of the training environment's physical representation of the real world and investigate policy architectures to bolster the robustness of the learned policy and minimizing the sim2real gap. To validate the proposed changes the trained policies will be deployed on embedded platforms such as PX4/Betaflight/etc. to generate low-level motor commands by on-device inference. An important challenge in this project is the development of a pipeline to enable on-device inference for the trained networks e.g. by decreasing the inference time of the network via network quantization. 
Jumping and in-flight-stabilization of a Jumping Quadruped in Mars Gravity Using Reinforcement Learning
Supervisors: Kostas Alexis, Jørgen Anker Olsen | Keywords: Legged Robots, Learning for Robotics​
Over the last decade's satellites, telescopes, landers, and wheeled rovers have been the main form of space exploration. As the field of legged robotics has developed and matured significantly in recent years, we now see the opportunity to explore more diverse and interesting terrain in space using specialized quadruped robots optimized for challenging off-world planetary environments, such as craters, caves, and lava tubes. Legged robots, such as the Boston Dynamics Spot and the ANYbotics ANYmal, present a set of advantages in mobility and versatility in complex environments over traditional wheeled robots and rovers. Jumping legged robots may be able to traverse the geometrically complex subterranean voids of lava tubes on planets such as Mars. A jumping legged robot for Martian surface and lava tube exploration will retain the key advantages of quadruped systems in overcoming rough terrain, while also being able to coordinate its actuators and exploit the low gravity environment of Mars and compliant leg designs to jump for significant height and thus overcome large obstacles. This project thesis aims to contribute to the modeling, control, and simulation of jumping legged robot that is currently being built and tested by our team at NTNU.   The main goal of the project thesis is to study and develop reinforcement learning techniques to enable the robot to perform mid-air stabilization using its legs. The second objective is to incorporate jumping and landing safely into this control policy. If progress is sufficient, deployment of the developed control policy to an actual legged robot may be possible. This continues work of previous master students with similar topic.​
Development and implementation of control policies for jumping and in-air-stabilization of a Jumping Quadruped in Mars Gravity
Supervisors: Kostas Alexis, Jørgen Anker Olsen | Keywords: Legged Robots, Learning for Robotics
Available. ​
Over the last decade, satellites, telescopes, landers, and wheeled rovers have been the main form of space exploration. As the field of legged robotics has developed and matured significantly in recent years, we now see the opportunity to explore more diverse and interesting terrain in space using specialized quadruped robots optimized for challenging off-world planetary environments, such as craters, caves, and lava tubes. Legged robots, such as the Boston Dynamics Spot and the ANYbotics ANYmal, present a set of advantages in mobility and versatility in complex environments over traditional wheeled robots and rovers. Jumping legged robots may be able to traverse the geometrically complex subterranean voids of lava tubes on planets such as Mars. A jumping legged robot for Martian surface and lava tube exploration will retain the key advantages of quadruped systems in overcoming rough terrain, while also being able to coordinate its actuators and exploit the low gravity environment of Mars and compliant leg designs to jump for significant height and thus overcome large obstacles. This project thesis aims to contribute to the modeling, control, and simulation of such a jumping legged robot that is currently being built and tested by our team at NTNU.   The main goal of the project thesis is to study, develop and implement classical and newer methods to enable the robot to perform mid-air stabilization using its legs, jumping, and walking. This continues the work of previous master students with similar topic.
Image from relevant project: https://www.spacehopper.ethz.ch/ ​
MPC-based Occlusion Avoidance for Object Monitoring
Supervisors: Kostas Alexis, Martin Jacquet | Keywords: Model Predictive Control
Available. ​
MPC have been used in the context of so-called perception aware applications, where vision-based constraints and objectives are integrated into the underlying optimization problem and solved in a receding horizon fashion. These approaches rely on the existence of a line of sight. However, this assumption is not easily verified in realistic scenarios, in uncontrolled and often cluttered environments. Occlusion avoidance is often tackled with geometric ray-casting to simulate the line of sight. But its evaluation is often computationally heavy, in particular when several occlusion obstacles are considered, and rely on a good knowledge of the environment. Moreover, these approaches rarely consider uncertainties associated with sensor measurements.  On the other hand, chance-constrained MPC are stochastic optimal control technics which have been successfully employed to tackle collision avoidance in receding horizon schemes. This thesis aims to address the problem of real-time occlusion avoidance in object monitoring in cluttered environments, exploiting stochastic constraints. 
Design, modeling and realization of a compliant swashplateless coaxial aerial vehicles colliding at high speeds
Supervisors: Kostas Alexis, Paolo De Petris | Keywords: Aerial Robotics
Available. ​
Coaxial aerial vehicles relying on swashplateless mechanism have proven to be able to emulate full actuation over forces and torques (six degrees of freedom) without a complex actuation design. Exploiting their smaller size, lightweight and controllability peculiarities, compared to standard quadcopters, this work focuses on the collision-tolerant properties of such design. The goal is to realize a compliant embodiment that permits the system to collide at very high speed without any repercussions.
Efficient Reinforcement Learning for Multirotor Design
​Supervisors: Kostas Alexis, Etor Arza | Keywords: Embodied Autonomy
Available. ​
Measuring the quality of the design of a multirotor involves testing the multirotor given a controller. While model based controllers can work for a variety of designs and don’t require any training, they are not always able to properly take advantage of the design. Learning based methods produce policies that better take advantage of the design, but they take a long time to train. This makes the co-optimization of design and control of multirotors a challenging optimization problem.
Safe Reinforcement Learning for visual Navigation using Control Barrier Functions
​Supervisors: Kostas Alexis, Marvin Harms | Keywords: Learning for Robotics
Available. ​
Autonomous robot navigation in cluttered environments poses an entire spectrum of challenges to any navigation stack. To overcome issues related to traditional navigation approaches using a full map to plan and execute actions, we consider a class of learned visual control policies to generate feasible actions based on exteroceptive sensor data such as depth images or LiDAR scans. Reinforcement Learning (RL) has successfully applied to such tasks previously [1], but the pertinent issue of safety of the learned control policy remains. In This work, we investigate the use of Control Barrier Functions within the training process as in [2] to guarantee safety of the learned policy for visual navigation tasks with high-dimensional input data. By using certificate functions such as CBFs to guide the training process, high levels of constraint satisfaction can be achieved. (references in the PDF)
Autonomous Ballast Tank Inspection with Underwater Robots
​Supervisors: Kostas Alexis, Mohit Singh | Keywords: Underwater Robotics
Available. ​
This project aims to develop an autonomous underwater robot for the task of autonomous exploration and inspection of ballast tanks in vessels (when filled with water). The system will be driven by vision and in this project we seek to develop/deploy visual-inertial SLAM, path planning for inspection/exploration and precise position control. On vision we will build upon recent contributions of our lab on refractive camera modeling and refraction-aware visual-inertial fusion. On path planning we will build on lab contributions originally built for flying and ground systems. On control we aim to rely primarily on open-source autopilots. The project focuses both on methods and their experimental validation. 
Semantic-oriented navigation of an Aerial Robot with Rotatable Camera
​Supervisors: Kostas Alexis, Marvin Harms | Keywords: perception, Semantics
Available. ​
Agile navigation is demonstrated using only a lightweight sensor suite like RGB-D cameras and IMU sensors. Nevertheless, existing works often utilize either visual or depth data coming from rigidly mounted cameras on the robotic platform. This data offers rich information about the surroundings in front of the robot only. Further, in most scenarios, the state-of-the-art navigation methods do not incorporate the semantic information from the RGB-D images. Thus, concerning semantic goal-oriented navigation, such a limited camera field-of-view leads to long operational time while the goal is to find a specific semantic object within the given environment. The integration of the rotatable camera within the navigation pipeline should significantly reduce the time along with the executed path by the robot. Simultaneously, it will allow for completing more missions/inspections within the given endurance time of the aerial platform. In this project, we aim to design and develop a rotatable camera and install it on the robotic system. In parallel, we strive to construct a novel semantic-oriented navigation pipeline which incorporates the rotatable camera.
Design, Development and Control of an Underwater Swimming Quadruped
​Supervisors: Kostas Alexis, Mohit Singh | Keywords: Underwater Robotics, Legged Robots
Available. ​
Legged robots go swimming! This project aims to develop a quadruped legged robot for underwater operations. The design shall present a host of gaits that allow the system to swim and also walk on the sea floor. Your task is to design the mechatronic solution, realize it in practice and develop control for its baseline leg gaits. 
Multi-modal SLAM Integration and Optimization for an Industrial Mobile Robot
​Supervisors: Kostas Alexis| Keywords: Ground Robotics
Available. ​
This project is in collaboration with Adigo Mechatronics and aims to deploy and optimize multi-modal (LiDAR, Vision, IMU) Simultaneous Localization And Mapping (SLAM) on one of the company's robot.
Modeling the Mechanisms of Human Visual Attention towards Objects of Interest
​Supervisors: Kostas Alexis| Keywords: Perception, Saliency
Available. ​
This project aims to model the mechanisms of human visual attention (bottom-up and top-down) towards objects of interest within the framework of inspection operations. To that end, this project shall exploit specialized glasses that capture the human gaze and localize where it points on a camera frame. By building a dataset appropriate for supervised learning a computational approach to model how inspectors attend and assess objects and structures of interest shall be developed.​

Own ideas - high risk projects!

Do you have your own idea about a robotics project? Are you willing to discuss a high-risk project with the understanding that things might not always work? Contact me and schedule a meeting!