Date of Graduation


Document Type


Degree Type



Statler College of Engineering and Mineral Resources


Mechanical and Aerospace Engineering

Committee Chair

Jason N. Gross

Committee Member

Guilherme A. S. Pereira

Committee Member

Marcello M. Napolitano

Committee Member

Natalia A. Schmid

Committee Member

Yu Gu


Robots’ autonomy has been studied for decades in different environments, but only recently, thanks to the advance in technology and interests, robots for underground exploration gained more attention. Due to the many challenges that any robot must face in such harsh environments, this remains an challenging and complex problem to solve.

As technology became cheaper and more accessible, the use of robots for underground ex- ploration increased. One of the main challenges is concerned with robot localization, which is not easily provided by any Global Navigation Services System (GNSS). Many developments have been achieved for indoor mobile ground robots, making them the easiest fit for subterranean explo- ration. With the commercialization of small drones, the potentials and benefits of aerial exploration increased along with challenges connected to their dynamics.

This dissertation presents two path planning algorithms for a team of robots composed of an Unmanned Ground Vehicle (UGV) and an Unmanned Aerial Vehicle (UAV) with the task of ex- ploring a subterranean environment. First, the UAV’s localization problem is addressed by fusing different sensors present on both robots in a centralized manner. Second, a path planning algo- rithm that minimizes the UAV’s localization error is proposed. The algorithm propagates the UAV motion model in the Belief Space, evaluating for potential exploration routes that optimize the sensors’ observations. Third, a new algorithm is presented, which results to be more robust to dif- ferent environmental conditions that could affect the sensor’s measurements. This last planning algorithm leverages the use of machine learning, in particular the Gaussian Process, to improve the algorithm’s knowledge of the surrounding environment pointing out when sensors provide poor observations. The algorithm utilizes real sensor measurements to learn and predict the UAV’s lo- calization error.

Extensive results are presented for the first two parts regarding the UAV’s localization and the path planning algorithm in the belief space. The localization algorithm is supported with real-world scenario experimental results, while the belief space planning algorithm has been extensively tested in a simulated environment. Finally, the last approach has also been tested in a simulated environ- ment and showed its benefits compared to the first planning algorithm.