Ywema, Yoes (2020) Learning from demonstration for isolating forest fires using convolutional neural networks. Bachelor's Thesis, Artificial Intelligence.
|
Text
AI_BA_2020_Yoes_Ywema.pdf Download (456kB) | Preview |
|
![[img]](https://fse.studenttheses.ub.rug.nl/style/images/fileicons/text.png) |
Text
toestemming.pdf Restricted to Registered users only Download (97kB) |
Abstract
Nowadays there are frequently occurring forest fires. A forest fire can be remedied by spraying water on it or digging firebreaks around it, so the fire cannot spread. The latter approach is studied in this thesis. With learning from demonstration a system is trained to dig firebreaks in a simulation with different forest fire environments. A deep convolutional neural network is used in this project to learn how to map environmental input to useful moves. The network is trained with different amounts of data and compared with a human interactive approach. In this human interactive approach the system asks a human to solve fires that it cannot isolate itself. The system can be used very effectively in simple environments but has to be expanded for more complex ones. In more complex environments a river and a number of houses result in a more complex sequence of steps to isolate the fire. Furthermore extra training data provides a way to improve most of the models that were previously trained with less data. Training models with extra training data, generated by the human interactive approach, resulted in the largest number of improved models.
| Item Type: | Thesis (Bachelor's Thesis) |
|---|---|
| Supervisor name: | Wiering, M.A. |
| Degree programme: | Artificial Intelligence |
| Thesis type: | Bachelor's Thesis |
| Language: | English |
| Date Deposited: | 30 Jan 2020 12:33 |
| Last Modified: | 31 Jan 2020 11:05 |
| URI: | https://fse.studenttheses.ub.rug.nl/id/eprint/21477 |
Actions (login required)
 |
View Item |