A neural network to classify auditory signals for use in autonomous harvester control systems
| dc.contributor.author | Simundsson, Avery | |
| dc.contributor.author | Thomas, Gabriel | |
| dc.contributor.author | Mann, Danny | |
| dc.contributor.examiningcommittee | Petkau, Donald (Biosystems Engineering) Thomas, Gabriel (Electrical and Computer Engineering) | en_US |
| dc.contributor.supervisor | Mann, Danny (Biosystems Engineering) | en_US |
| dc.date.accessioned | 2019-08-29T17:11:36Z | |
| dc.date.available | 2019-08-29T17:11:36Z | |
| dc.date.issued | 2019-07-22 | en_US |
| dc.date.submitted | 2019-08-17T14:51:01Z | en |
| dc.degree.discipline | Biosystems Engineering | en_US |
| dc.degree.level | Master of Science (M.Sc.) | en_US |
| dc.description.abstract | 3 1 ABSTRACT As agricultural machinery moves into the digital era, significant developments in the available technology make autonomous farm vehicles more feasible, affordable, and desirable. One of the challenges of effective autonomous vehicle control specific to agriculture is the ability of the vehicle to interpret and adapt to constantly changing conditions. There are many types of sensors able to identify specific changes in conditions (elevation, temperature, image etc.), but a single indicator to signal a variety of changes in operating conditions would be beneficial to a remote human operator, particularly in triggering an automatic shutdown to prevent machinery damage. Auditory information is a primary indicator of changing conditions to an in-cab operator, particularly in detecting mechanical overload in a combine. This paper explores the potential for auditory information, which has proven valuable to an in-cab operator, to be used in autonomous vehicle control. Sound was recorded at a sampling rate of 48 kHz near the combine chopper for three different operating modes during the same harvest day for canola. Samples from each clip were segmented and analyzed using the Fast Fourier Transform (FFT) in MATLAB. The FFT generated a power spectral density (PSD) function for each segment, from which eight features were extracted and labelled to create feature vectors. These vectors were used to create a classifier using a feedforward pattern recognition neural network. The network used scaled conjugate gradient backpropagation training to achieve accuracy of 100%. The speed of sampling and analysis is sufficient to be used in real time machinery analysis and control. | en_US |
| dc.description.note | October 2019 | en_US |
| dc.identifier.uri | http://hdl.handle.net/1993/34105 | |
| dc.language.iso | eng | en_US |
| dc.rights | open access | en_US |
| dc.subject | Autonomous, FFT, Machine Learning, Control system, Auditory, Sonifications | en_US |
| dc.title | A neural network to classify auditory signals for use in autonomous harvester control systems | en_US |
| dc.type | master thesis | en_US |