Steering Angle Prediction by a Deep Neural Network and its Domain Adaption Ability
(2018) In Master's Theses in Mathematical Sciences FMAM05 20172Mathematics (Faculty of Engineering)
- Abstract
- The goal of this thesis is to design an artificial neural network for self-driving vehicles in regards to steering the vehicle. The performance of the networks is evaluated on a simulated race track in addition to more conventional metrics such as cost/loss. In the main part of the project the networks were trained on data from real-life driving and validated/tested on simulated data, which is an example of domain adaption. The simulated data were from the same conditions as in the simulated race track. Two main designs were evaluated, one based on the design proposed by NVIDIA and one based on the idea of multi-task learning where an autoencoder was trained simultaneously with a steering angle predictor.
The resulting network based... (More) - The goal of this thesis is to design an artificial neural network for self-driving vehicles in regards to steering the vehicle. The performance of the networks is evaluated on a simulated race track in addition to more conventional metrics such as cost/loss. In the main part of the project the networks were trained on data from real-life driving and validated/tested on simulated data, which is an example of domain adaption. The simulated data were from the same conditions as in the simulated race track. Two main designs were evaluated, one based on the design proposed by NVIDIA and one based on the idea of multi-task learning where an autoencoder was trained simultaneously with a steering angle predictor.
The resulting network based on the multi-task learning and solely trained on real-life driving data managed to make it around the entire simulated track without driving off the road. The network based on the NVIDIA design on the other hand only managed to stay on the road for a short period of time under the same conditions. The results indicate that the multi-task based design is better at domain adaption than the one based on NVIDIA's design and incites for further research within the area.
Also some additional tests were conducted with real-life data for both training and validation. The results from this part of the thesis were ambiguous with respect to the domain adaption ability. (Less) - Popular Abstract
- In recent years the development of self driving cars has seen some great progress. More and more companies and organizations have vehicles out in traffic that partly steer themselves. However, the process of reaching a level where a car can drive satisfactory by its own is a long one and involves a lot of training. Until today, the most common approach is to expose the system to as many different situations as possible as it may encounter in reality. All in order to make the car as versatile as possible. The mathematical system responsible for the steering is called an artificial neural network. A much wanted feature in such a system is the ability to generalize and use prior knowledge in new situations. In the same way as humans can tell... (More)
- In recent years the development of self driving cars has seen some great progress. More and more companies and organizations have vehicles out in traffic that partly steer themselves. However, the process of reaching a level where a car can drive satisfactory by its own is a long one and involves a lot of training. Until today, the most common approach is to expose the system to as many different situations as possible as it may encounter in reality. All in order to make the car as versatile as possible. The mathematical system responsible for the steering is called an artificial neural network. A much wanted feature in such a system is the ability to generalize and use prior knowledge in new situations. In the same way as humans can tell that a golden retriever is a dog, despite never seeing a golden retriever before. In a recent master's thesis from Lund University, promising results are presented for artificial neural networks for self-steering cars. The results show that certain features of a network may increase its ability to generalize. (Less)
Please use this url to cite or link to this publication:
http://lup.lub.lu.se/student-papers/record/8946239
- author
- Hansson, Anders LU
- supervisor
- organization
- course
- FMAM05 20172
- year
- 2018
- type
- H2 - Master's Degree (Two Years)
- subject
- keywords
- machine learning, domain adaption, self-steering vehicles, neural networks, convolutional neural networks, artificial neural network, supervised learning
- publication/series
- Master's Theses in Mathematical Sciences
- report number
- LUTFMA-3349-2018
- ISSN
- 1404-6342
- other publication id
- 2018:E26
- language
- English
- id
- 8946239
- date added to LUP
- 2018-06-12 16:49:55
- date last changed
- 2018-06-12 16:49:55
@misc{8946239,
abstract = {{The goal of this thesis is to design an artificial neural network for self-driving vehicles in regards to steering the vehicle. The performance of the networks is evaluated on a simulated race track in addition to more conventional metrics such as cost/loss. In the main part of the project the networks were trained on data from real-life driving and validated/tested on simulated data, which is an example of domain adaption. The simulated data were from the same conditions as in the simulated race track. Two main designs were evaluated, one based on the design proposed by NVIDIA and one based on the idea of multi-task learning where an autoencoder was trained simultaneously with a steering angle predictor.
The resulting network based on the multi-task learning and solely trained on real-life driving data managed to make it around the entire simulated track without driving off the road. The network based on the NVIDIA design on the other hand only managed to stay on the road for a short period of time under the same conditions. The results indicate that the multi-task based design is better at domain adaption than the one based on NVIDIA's design and incites for further research within the area.
Also some additional tests were conducted with real-life data for both training and validation. The results from this part of the thesis were ambiguous with respect to the domain adaption ability.}},
author = {{Hansson, Anders}},
issn = {{1404-6342}},
language = {{eng}},
note = {{Student Paper}},
series = {{Master's Theses in Mathematical Sciences}},
title = {{Steering Angle Prediction by a Deep Neural Network and its Domain Adaption Ability}},
year = {{2018}},
}