Use this URL to cite or link to this record in EThOS:
Title: Robot mapping and localisation in water pipes
Author: Ma, Ke
ISNI:       0000 0004 7431 1506
Awarding Body: University of Sheffield
Current Institution: University of Sheffield
Date of Award: 2018
Availability of Full Text:
Access from EThOS:
Access from Institution:
The demand for inspection and repair technologies for the water industries on their water mains and distribution pipes is increasing. In urban water distribution systems, due to the fact that water pipes are ageing, pipe leakages and serious damage may occur. Compared with the cost of pipe replacement in the underground distribution system, regular pipe inspection and repair is more cost effective to water companies and local communities. However, small-diameter pipes are not accessible to humans because they are small in size and often buried underground. Therefore, inspection robotic systems are more suited to this task in terms of underground pipe networks mapping and damage localisation, in order to target repair from above ground. There are a number of challenges for robot mapping and localisation in water pipes, which are: 1) feature sparsity in water pipes – lack of navigation landmarks, 2) in-pipe robot can only detect nearby features, and 3) unlike indoor/outdoor SLAM problems, in-pipe robot has less movement flexibility. The main aim of this thesis is to solve these challenges and address the problem of robot mapping and localisation in small-diameter feature-sparse water pipes. This thesis presents a number of novel contributions. Firstly, for the front end, where raw sensor data is transformed into signals useful for mapping and localisation algorithms, new types of maps are developed here for water pipes: for plastic pipes, ultrasound data is used to map the ground profile through the plastic pipe wall, whilst for metal pipes a hydrophone is used to determine a map based on pipe vibration amplitude over space. Secondly, a new sequential approach to mapping and localisation is developed, based on alignment of multiple maps based on dynamic time warping averaging. Thirdly, a new simultaneous localisation and mapping algorithm is developed, which overcomes the limitation of the sequential approach in that the map is not updated. Finally, a new sensor fusion algorithm is developed that transforms the robot location in the local coordinate frame to the world coordinate frame, which would be essential for targeting repairs from above ground.
Supervisor: Anderson, Sean ; Richard, Collins ; Tony, Todd Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available