Isherwood, Zach and Secco, Emanuele Lindo (2021) A Raspberry Pi computer vision system for self-driving cars. In: Computing Conference 2022, 14-15 July 2022, London. (Accepted for Publication)
Preview |
Text
Zach_paper.pdf - Accepted Version Download (690kB) | Preview |
Abstract
This paper presents a prototype of a self-driving vehicle that can detect the lane that it is currently in and can aim to maintain a central position within that lane; this is to be done without the use of special sensors or devices and utilizing only a low-cost camera and processing unit. The proposed system uses a hand-built detection system to observe the lane markings using computer vision, then using these given lines, calculate the trajectory to the center of the lane. After locating the center of the lane, the system provides the steering heading that the vehicle needs to maintain to continuously self-correct itself; this process is real-time performed with a sampling frequency of 20 Hz. Due to the increased number of calculations, the heading is smoothed to remove any anomalies in observations made by the system. Since this system is a prototype, the required processing power used in an actual vehicle for this application would be much higher since the budget of the components would be more significant; a higher processing speed would lead to an overall increased frame rate of the system. In addition, a higher frame rate would be required for higher speeds of the vehicle to allow for an accurate and smooth calculation of heading. The prototype is fully operational within an urban environment where road markings are fully and clearly defined along with well-lit and smooth road surfaces.
| Item Type: | Conference or Workshop Item (Paper) |
|---|---|
| Additional Information and Comments: | This version of the paper has been accepted for presentation at Computing Conference 2022. |
| Faculty / Department: | Faculty of Human and Digital Sciences > Mathematics and Computer Science |
| Depositing User: | Emanuele Secco |
| Date Deposited: | 12 Nov 2021 11:56 |
| Last Modified: | 15 Jul 2022 00:15 |
| URI: | https://hira.hope.ac.uk/id/eprint/3421 |
Actions (login required)
 |
View Item |
Altmetric
Altmetric
