Options
Bukhari Ilias
Preferred name
Bukhari Ilias
Official Name
Bukhari, Ilias
Alternative Name
Ilias, Bukhari
Ilias, B.
Main Affiliation
Scopus Author ID
55362012800
Researcher ID
EZX-8508-2022
Now showing
1 - 1 of 1
-
PublicationEnhancement of ultrasonic sensor system for indoor mobile robot mapping and self-localizationThis thesis is about enhancing the ability of ultrasonic sensor for accurate measurement of obstacles, which is important in indoor mobile robot mapping and self-localization application. Ultrasonic is preferred for indoor mobile robot mapping and selflocalization due to its wide detection area, does not depends on light, the ability to detect glass and shining wall, smaller in size, lightweight, use a very low memory, much cheaper than Laser Range Finder (LRF) or camera and lower power consumption. However, it can lead towards inaccurate measurement due to its characteristics. Being a specular type of sensor and not narrowly focused, this can causes wrong measurement of estimation after multiple obstacle reflection. Therefore, necessary accompanying algorithms should be developed to encounter this issue, which lead towards this research that consists of three objectives, based on theoretical and experimental work. The first is the development of an enhanced wall mapping technique named as the Nominal Wall Angle (NWA) filtering algorithm, using the combination of Sine and Cosine Law. The second is the development of real time 2D mapping based on the measurement distance between the centre of sensor bank and the wall. Homogeneous Transformation Matrix (HTM) is enhanced in this research as the wall mapping algorithm. The third objective is to evaluate the performance of mapping and selflocalization of basic wall shape and real indoor environment using K-Nearest Neighbour (KNN) classifier. To achieve these objectives, a ring of sixteen pieces 40 kHz ultrasonic sensor bank is developed. This sensor bank is mounted on a two-wheeled mobile robot platform to perform the scanning process. The robot platform was built using a differential drive system as it can handle issues of maneuvering in crowded and narrow spaces. Analysis for enhancing ultrasonic capability in obstacle measurement with ultrasonic sensor bank is carried out by fixing four set of basic walls shape of rectangular, triangular, curve and square and also in a real laboratory environment. Normality and Multivariate analysis of variance (MANOVA) test was performed to ensure that the data is valid. The algorithm has the combined implementation of several software platforms such as BASIC Stamp, BASIC Atom, Lab VIEW and MA TLAB. The combination of these algorithms, hardware and software results in an improvement in accuracy of mapping, increased from an average of 90% to 97%. Also at the same time, data dimension reduction of up to 75% can be achieved while maintaining localization performance rate beyond 90%. The developed algorithm has the capability to suppress phantom error for improving the accuracy of measurements during the scanning process.